On the developmental signifi cance of female pubertal timing

On the developmental signifi cance of female pubertal timing
On the developmental significance
of female pubertal timing
This dissertation is dedicated
to my mother, Birgitha, my sister, Sofi a,
and the memory of my father, Lennart
“Bryt upp, bryt upp! Den nya dagen gryr.
Oändligt är vårt stora äventyr”.
karin boye
“…for right timing is in all things
the most important factor.”
hesiod
Örebro Studies in Psychology 14
Therése Skoog
On the developmental significance
of female pubertal timing
© Therése Skoog, 2008
Title: On the developmental significance of
female pubertal timing
Publisher: Örebro University 2008
www.publications.oru.se
Editor: Heinz Merten
[email protected]
Printer: Intellecta DocuSys, V Frölunda 03/2008
issn 1651-1328
isbn 978-91-7668-590-7
Abstract
Skoog, T. (2008) On the developmental significance of female pubertal timing. Örebro
Studies in Psychology 14.
Puberty is the process of becoming sexually mature and it has fundamental somatic and
psychosocial implications. The focus of this dissertation is on the short- and long-term
developmental significance, concerning soma et psyche, of female pubertal timing. Four
studies were designed and performed to accomplish these aims. Six samples of different ages
from different countries and at different time points, comprising several thousand females,
some of whom were followed longitudinally, were used. Age at menarche was used as the
measure of pubertal maturation. The first main aim of this dissertation was to explore the
mechanisms that might explain the well-established link between female pubertal timing and
problem behavior and to identify the contextual conditions under which associations are
stronger or weaker. Existing explanations are unsatisfactory, and little is known about
conditions that might affect the strength of the associations.
For Paper I, we tested and confirmed a peer socialization hypothesis as a satisfactory
explanation for the link between early puberty and problematic adjustment. In short, this
hypothesis posits that early-developing girls associate with older peers and boyfriends
because they feel more mature than their same-age peers and – through these peers and
boyfriends – are channeled into more socially advanced behaviors, including problem
behavior. This should be particularly true in contexts where heterosexual relationships are
sanctioned and where there is an abundance of deviant youth. For Paper II, I used a
biopsychosocial approach, and investigated pubertal timing along with self-perceptions of
maturity and early romantic relationships. The findings revealed that early puberty had very
different implications depending on the psychological and social contexts in which it was
embedded. For instance, when early puberty was coupled with feeling mature and having
early romantic relationships, it was associated with adjustment problems. When early puberty
was coupled with neither, it was not linked to particularly high levels of problem behavior.
In stark contrast to the vast literature on the role of female pubertal timing in adolescence,
the literature on long-term implications is remarkably limited. For this reason, the second
main aim of this dissertation was to study the adult implications of female pubertal timing.
For papers III and IV, we examined the long-term implications of pubertal timing, particularly
as it relates to somatic development. The findings suggest that pubertal timing does have
future implications for women’s body perception and morphology, with early-developing
females having higher body mass in adulthood, but only under certain circumstances. The
findings of this dissertation help further understanding of the soma et psyche implications of
female pubertal timing. They indicate that pubertal timing has concurrent and future
implications. It seems, however, that timing is not everything. The developmental significance
of female pubertal timing appears to be very different under different contextual conditions.
Thus, it is only when girls’ psychological and social contexts are considered that fruitful
predictions can be made. As such, the findings have important implications for prevention,
policy and practice.
Keywords: female, puberty, pubertal timing, development, adjustment, longitudinal study,
sexuality, peer relations, weight status, mechanisms, conditions
[email protected]
Acknowledgements
I would like to express my sincere gratitude to all those who supported and inspired me
throughout my doctoral training. Writing this dissertation, I have had the pleasure of working
on several unique longitudinal research projects. For this I am forever thankful. Several
papers in this dissertation were made possible by access to data from the longitudinal research
program Individual Development and Adaptation (IDA). The scientific leader is Lars R.
Bergman. Responsible for the planning, implementation and financing of the collection of
data before 1996 was David Magnusson and after that time Lars R. Bergman. The data
collections were supported by grants from the Swedish National Board of Eduction, the
Swedish Committee for the Planning and Coordination of Research, The Bank of Sweden
Tercentenary Foundation, the Swedish Social Research Council. I would like to thank
Professor David Magnusson, the founder of IDA, and Professor Lars R Bergman at
Stockholm University for letting me use data from the IDA project. I would like to thank
Professor Håkan Stattin and Professor Margaret Kerr for letting me work with the other data
used in this dissertation, including those from the Solna and the 10 to 18 studies. Most
importantly, I would like to express my sincere gratitude to all the people who took part in the
studies in this dissertation. Without you, there would have been nothing to write about.
I have had the opportunity to have Professor Håkan Stattin as my supervisor. As a doctoral
student, you are supposed to learn about research. You taught me more than that. You taught
me how to be a researcher. As I see it, this is a far more important lesson. You have guided
me through the work for this dissertation, provided a rich working environment, and
introduced me to people who later have become close collaborators. Most importantly, you
have taught me that doing research could actually be fun! Being a sensation seeker, this lesson
has been the most crucial of all.
Professor Margaret Kerr has been my co-supervisor. You have always taken the time to
answer and discuss the questions I have had. It is has been truly inspiring to work with such a
successful researcher. Thanks to both you and Håkan for giving me the opportunity to be part
of your research group.
The Center for Developmental Research is the research group with which I have been
affiliated. It is a research organization where people actually work together as a group and the
group members support each other, read each others manuscripts, and deliver fruitful
comments. Being a member of this research group is inspiring, educative, and fun. I would
like to thank all fellow doctoral students and post docs who are or who have been working at
the Center, Andreas, Bill, Fumiko, Gowert, Jenny, Luna, Maria, Mats, Michael, Nejra,
Nanette, Nikolaus, Stefan, Terese, Vilmante, Vivi, and Ylva for the time we have worked
together. I am particularly grateful to Vivi for her excellent assistance with the other research
projects that I have worked on while finishing this dissertation. I am also grateful to Camilla
for professional help with teaching.
I would like to thank Associate Professor Maria Tillfors, the internal reviewer, and Professor
Lars Wichström, the external reviewer, for insightful and valuable comments on an earlier
version of this dissertation.
I am sincerely grateful to Jon Kimber for carefully editing my English and to Birgitta Kimber
for lots of support.
When I started to work at the Department of Social Sciences in 2002, Louise Svensson, who
is now a doctoral student in Sociology at Örebro University, introduced me to working in
academia. Ever since, you have been an infinite source of support, professionally and
privately. You are one of the wisest, not to mention strongest, women I know. It is a blessing
having you as a friend.
I would also like to thank my close and very dear friends, Helena and Joakim Lövman, for
unceasingly having put up with my “Research shows…” comments. I am looking forward to
an endless array of further adventures with both of you in the future. Helena, considering the
rocket career that you have done (going from a part-time salary and wage administrator to
being a human-resource manager with responsibility for several hundred employees in less
than a year), I would very much like to get your work-life advice now that I am finally done
with being a student.
Henrik, you are a true friend. During the last decade, you have been the greatest source of
support I have had. Thank you so much for everything. I am sure you will make an excellent
police officer in a very nearby future.
I would like to thank my mother, Birgitha, to whom I have also dedicated this dissertation.
There is an obvious reason why I would not have written this dissertation if it was not for you,
but there are many others too. You have taught me to always believe in what I do, and you
have told me that the sky is the limit for what I, or anyone else, can achieve. Thanks my sister,
Sofia, who has brought me back to reality at times when I have been most disconnected from
the real world. Jag vill även ge ett stort tack till min mormor Sonia för att du finns!
Finally, I would like to thank Per, my wonderful loving husband, for his infinite support and
endless intellectual challenges. Words are not enough to describe how I feel about you and
how grateful I am for being able to share my life with you. Thanks also to our child, who has
not yet been born, who appeared at a time in my life (i.e., when trying to finish this
dissertation) when I truly needed to be reminded about what matters the most.
Stora Brandsjöhults gård, March 2008
List of papers
This dissertation is based on the following papers, which hereafter will
be referred to by their Roman numerals.
Paper I
Kerr, M., Skoog, T., Stattin, H., & Ruiselova, Z. (2008).
Female Pubertal Timing and Problem Behavior: Explaining
the Mechanism at Different Levels of Social Contexts.
Manuscript.
Paper II Skoog, T. (in press). Concurrent and future implications of girls’
pubertal timing: What roles do perceived maturity and early
sexuality play? In A.-K. Andershed & C. Magnusson (Eds.),
Troubled girls and girls’ troubles. New York: Springer.
Paper III Johansson, T., & Ritzén, M. (2005). Very long-term follow-up
of girls with early and late menarche Endocrine Development, 8,
126–136.
Paper IV Skoog, T., & Stattin, H. (2008). Does appetite affect
the association between female pubertal timing and adult
weight status?
Manuscript to be resubmitted for publication.
Paper III has been reprinted with the kind permission of S. Karger AG,
Basel.
Table of contents
1. Introduction .................................................................................... 15
1.1 The developing body ....................................................................... 16
1.1.1 Endocrine changes ................................................................ 16
1.1.2 Maturation of primary sexual characteristics.......................17
1.1.3 Appearance of secondary sexual characteristics ...................17
1.1.4 Growth spurt ........................................................................17
1.2 Pubertal status and pubertal timing ............................................... 18
1.3 Measuring puberty ......................................................................... 19
1.4 Adolescent implications of pubertal timing ....................................22
1.4.1 Somatic implications ............................................................22
1.4.2 Psychosocial implications .....................................................23
1.5 Adult implications of pubertal timing .............................................30
1.5.1 Somatic implications ............................................................30
1.5.2 Psychosocial implications .....................................................30
1.6 Explaining the link between pubertal timing and adjustment ........32
1.7 Remaining questions – limitations of the existing literature ...........34
2. Aims of the dissertation .................................................................. 37
2.1 The biopsychosocial approach ........................................................ 37
2.2 The peer socialization hypothesis .................................................. 38
2.2.1 The role of contexts ..............................................................40
2.3 Purpose of the papers in the dissertation ........................................41
3. Method ............................................................................................ 43
3.1 Participants and procedures ............................................................ 43
3.1.1 Sample I ................................................................................. 43
3.1.2 Sample II ................................................................................44
3.1.3 Sample III ............................................................................... 45
3.1.4 Sample IV .............................................................................. 45
3.1.5 Sample V ................................................................................ 45
3.1.6 Sample VI ..............................................................................46
3.2 Measures ........................................................................................49
3.2.1 Paper I....................................................................................49
3.2.2 Paper II .................................................................................. 55
3.2.3 Paper III .................................................................................57
3.2.4 Paper IV................................................................................. 58
3.3 Statistical analyses ..........................................................................59
4. Results ............................................................................................. 61
4.1 Paper I ............................................................................................. 61
4.1.1 The research question of Study 1 .......................................... 61
4.1.2 The fi ndings of Study 1......................................................... 61
4.1.3 The conclusion of Study 1 .................................................... 61
4.1.4 The research question of Study 2 ..........................................62
4.1.5 The fi ndings of Study 2.........................................................62
4.1.6 The conclusion of Study 2 ....................................................62
4.1.7 The research question of Study 3 ..........................................62
4.1.8 The fi ndings of Study 3.........................................................62
4.1.9 The conclusion of Study 3 ....................................................64
4.1.10 The research question of Study 4 ..........................................64
4.1.11 The fi ndings of Study 4.........................................................64
4.1.12 The conclusion of Study 4 .................................................... 65
4.2 Paper II ............................................................................................ 65
4.2.1 The research question of Paper II ......................................... 65
4.2.2 The fi ndings of Paper II ........................................................ 65
4.2.3 The conclusion of Paper II ....................................................67
4.3 Paper III ...........................................................................................67
4.3.1 The research question of Paper III ........................................67
4.3.2 The fi ndings of Paper III .......................................................67
4.3.3 The conclusion of Paper III ...................................................68
4.4 Paper IV ...........................................................................................68
4.4.1 The research question of Paper IV ........................................68
4.4.2 The fi ndings of Paper IV .......................................................68
4.4.3 The conclusion of Paper IV...................................................69
5. General discussion ...........................................................................71
5.1 The fi ndings in the four papers and how they
relate to previous research ...............................................................71
5.1.1 Support for the peer socialization hypothesis ........................71
5.1.2 Combining the peer socialization and the contextual
amplification hypotheses ................................................................. 74
5.2 The papers as a whole – the biopsychosocial approach ....................78
5.3 Methodological limitations and strengths ......................................80
5.4 Implications for policy, prevention, and practice ............................. 82
5.5 Future research directions ...............................................................84
5.5.1 Changing societies – does research keep up? .........................86
5.6 Summary and fi nal remarks.............................................................87
5.7 Conclusions .....................................................................................89
References ............................................................................................. 91
1
1. Introduction
Eight hundred years before Christ, Hesiod, a Greek didactic poet, coined the phrase “Observe
due measure, for right timing is in all things the most important factor”. Today, we would say
“Timing is everything”. For many years, this expression has been widely used in as different
contexts as car sales, cancer therapy, and everyday language. According to the existing
research literature, it can be applied to the developmental significance of female puberty as
well (Pinyerd & Zipf, 2005).
Adolescence is a period of dramatic emotional, cognitive, social, and biological change
(Patton & Viner, 2007). It takes place after what is one of the longest childhood periods of all
mammals (Grumbach & Styne, 2003). Some of the most fundamental changes in adolescence
are biological, and the biological changes associated with puberty are commonly regarded as
a signal of the onset of adolescence (Petersen, 1998). Puberty refers to the biological changes
that are needed for sexual maturity, which all healthy individuals experience. With this
transition, girls go from having the appearance of children to resemble the appearance of the
adult female stereotype (Tanner, 1978). The term puberty stems from pubertas, which means
adult in Latin, or pubescere, which means growing hairy. The Merriam-Webster online
dictionary (2007) defines puberty as “the condition of being or the period of becoming first
capable of reproducing sexually marked by maturing of the genital organs, development of
secondary sex characteristics, and in the human and in higher primates by the first occurrence
of menstruation in the female”. Thus, the essence of female puberty is sexual maturation.
Researchers and the public alike have viewed puberty as a period of “sturm und drang”
ever since the beginning of the 20th century, and long before that too (e.g., Arnett, 1999;
Buchanan, Eccles, Flanagan, Midgley, Fiedlaufer, & Harold, 1990; Hall, 1904).
Contemporary researchers (see Arnett, 1999) still argue that adolescents are more likely to
experience problems than any other age group. Peripubertal youth are seen as difficult to deal
with, and people tend to think that temper tantrums are part of youths’ everyday lives (e.g.,
Holmbeck & Hill, 1988). Fifteen years ago, Buchanan and colleagues (Buchanan, Eccles, &
Becker, 1992) concluded, in their review of the literature on the role of the endocrinological
aspects of puberty for behavior, that radical hormonal changes during puberty contribute to
emotional volatility. Furthermore, there are findings suggesting that puberty does not only
affect the developing person, but also his or her parents. When a child (particularly the
firstborn) is experiencing puberty, the parents experience changes in marital qualities,
showing less positivity and more negativity (Whiteman, McHale, & Crouter, 2007). Thus,
puberty has long been regarded as meaning trouble.
There is, however, evidence to suggest that this might not always be true. Somewhat
contrary to common belief and the conclusion drawn by Buchanan et al. (1992), Larson and
Richards (1994) showed that there was little connection between mood disruptions and
pubertal stage (i.e. where in the pubertal process the youth was). Furthermore, others have
15
15
2
argued that experiencing storm and stress is neither universal nor inevitable during this period
in life (Arnett, 1999). So, if not all adolescents experience problems during puberty, who
will?
The answer seems to involve the timing of the pubertal changes. Research in the
adolescent adjustment area suggests that it is rather when (in relation to peers) girls
experience puberty – than puberty per se – that is of particular significance for adolescent
functioning (Weichold, Pröhl, Büttig, & Silbereisen, 2007). In other words, research has
indicated that “timing seems to be everything”. A recent publication has even been called
“Puberty-Timing is everything!” (Pinyerd & Zipf, 2005). This seems to be true for somatic
aspects, such as weight status and internalizing and externalizing behaviors, and also for
substance use, school adjustment and social relationships (e.g., Alsaker, 1995, 1996;
Buchanan et al., 1992; Celio, Karnik, & Steiner, 2006; Graber, 2003; Mendle, Turkheimer, &
Emery, 2007; Stattin & Magnusson, 1990; Susman, Dorn, & Schiefelbein, 2003; Susman &
Rogol, 2004; Waylen & Wolke, 2004, for reviews). In all these areas, it is girls who mature
early who are at heightened risk of experiencing difficulties. Why this is and whether pubertal
timing continues to play a role in adulthood are largely unknown.
To better understand its developmental significance, I place female puberty within a
developmental perspective. The main aim is to investigate the short and long-term
developmental significance, concerning soma et psyche, of female pubertal timing. In short, I
investigate somatic as well as psychosocial implications of pubertal timing, and I focus on the
long-term implications and mechanisms or conditions that might explain its impact. While
early puberty puts girls at higher risk of negative outcomes, far from all early maturers have
problems. Therefore, I am interested in why some but not other early-developed girls
experience difficulties and what other factors may be involved.
1.1 The developing body
Puberty is a biological phenomenon, with psychological and social meaning. Knowledge of
the physiological processes of puberty is needed for understanding how puberty might
influence development. Puberty comprises several processes, each of which might have
different developmental implications. The following section gives a brief description of
pubertal changes.
1.1.1 Endocrine changes. Puberty is a phase of anatomical and physiological
development during which the sex organs mature and become functional. There is a
developmental continuum that starts during the fetal period, goes through puberty, and ends
with full sexual maturity (Grumbach & Styne, 2003). Pubertal changes can all be related to
the endocrine, or hormonal, system, particularly to the hypothalamic-pituitary-adrenal (HPA)
and hypothalamic-pituitary-gonadal (HPG) axes. Adrenarche refers to the maturation of the
adrenal gland, which is regulated by the HPA axis. This activation results in increased
concentration of adrenal androgens, which takes place during the first phases of puberty.
16
16
3
These androgens, in turn, results in body odor, pubic hair, and often oily skin, including acne.
Adrenarche can start at the age of six in females and continue into their thirties (Dorn &
Rotenstein, 2004; Grumbach & Styne, 2003). Gonadarche refers to the maturation of the
ovaries (gonads). This results in the stimulation of the ovaries, producing sex hormones
(estradiol in girls). Gonadarche is caused by activation of the HPG axis. It is related to
maturation of the primary sexual characteristics, including menarche, and the development of
secondary sexual characteristics (Dorn, Dahl, Woodward, & Biro, 2006). Gonadrache begins
around age nine or ten on average (Grumbach & Styne, 2003). It indicates that puberty has
started. Adrenarche and gonadarche are the two main endocrinological processes during
puberty.
1.1.2 Maturation of primary sexual characteristics. The primary sexual
characteristics are physical characteristics of the reproductive system that mature during
puberty. The sex organs grow in size and they start to produce sex cells. Girls’ menarche (i.e.,
their first menstrual period) is related to these changes. Menarche is one of the milestones in
female development, and is often taken to indicate fertility. However, it can take another
several years before the adolescent girl reaches full sexual maturity. On average, menarche
occurs at age 12.6 years in white girls in Western countries (Biro, Huang, Crawford, Lucky,
Striegel-Moore, Barton et al., 2006). It occurs in middle or late puberty, about two years after
the onset of breast development (Pinyerd & Zipf, 2005). Overall, the purpose of the
anatomical and physiological development of the primary sexual characteristics is to make the
individual capable of reproduction and to transform a sexually immature girl into a sexually
mature woman.
1.1.3 Appearance of secondary sexual characteristics. Secondary sexual
characteristics are not directly associated with reproduction. They are what make girls look
more similar to the adult female stereotype, and they reflect sexual dimorphism (i.e., sexdetermined physical differences). As such, the secondary sexual characteristics signal sexual
maturity and reproductive capacity to others. In girls, breast development is one of these
changes; the addition of body fat, particularly to the hips, is another. Estradiol primarily
targets breast tissue, and the development of breast buds is normally the first sign of puberty
in girls (van den Berg, Setiawan, Bartels, Polderman, van der Vaart, & Boomsma, 2006),
which occurs approximately two years before menarche (Patton & Viner, 2007).
1.1.4 Growth spurt. In addition to changes in primary and secondary sexual
characteristics, a rapid growth spurt takes place that results in adult height and weight. This
growth spurt lasts approximately two years (Abassi, 1998). Mean age at peak height velocity
is around 11 years in North American girls, and girls’ heights increase around 25-30
centimeters during this period (Abassi, 1998).
Peak weight velocity takes place approximately six months after peak height velocity
(Rogol, Roemmich, & Clark, 2002). During this period, there is increased fat and muscle
growth. The fat contributes to the typical female hourglass shape as it moves fat from the
17
17
4
middle to the upper and lower parts of the body. Puberty also leads to increased strength
and endurance, partly because of musculoskeletal development and partly because of
maturation of the respiratory and circulatory organs. These rapid changes in girls’
statures are easily noticed by the adolescents themselves, and also people in their
surroundings. Together with breast development, acceleration in growth velocity is one
of the first overt signs of pubertal maturation in girls (Grumbach & Styne, 2003; Parent,
Teilmann, Juul, Skakkebaek, Toppari, & Bourguignon, 2003).
1.2 Pubertal timing and pubertal status
Social scientists have generally paid attention to two aspects of pubertal maturation,
pubertal status and pubertal timing (Costello, Sung, Worthman, & Angold, 2007).
Pubertal status refers to the level of maturation a girl has reached at a given point in time.
Research on the role of pubertal status has, for instance, examined the influence puberty
has on girls’ moods (Buchanan et al., 1992). When examining pubertal status, researchers
are interested in the linear effects of pubertal maturation (i.e., if the chance of a given
outcome increases, or decreases, as a function of maturation) or how the level of pubertal
development impacts on a certain outcome. A body of research has shown, however, that
the timing of pubertal changes is more influential than the changes per se for
psychosocial adjustment, and also for several physiological factors, such as overweight,
breast cancer, and osteoporosis (e.g., Alsaker, 1996; Biro, McMahon, Striegel-Moore,
Crawford, Obarzanek, Morrison et al., 2001; Blum, Harris, Must, Phillips, Rand, &
Dawson-Hughes, 2001; Brooks-Gunn, Petersen, & Eichorn, 1985; Freedman, Khan,
Serdula, Dietz, Srinivasan, & Berenson, 2003; Rockhill, Moorman, & Newman, 1998).
Although puberty is a universal transition, girls differ markedly with respect to how
old they are when they experience it. Pubertal timing refers to when – in relation to samesex, same-age peers girls experience puberty (Costello et al., 2007) – and girls can differ
from each other by many years with respect to when they go through puberty. Age at
menarche can vary from around 8 to 17 in healthy white girls (Tremblay & Frigon,
2005). Everything up to 2.5 standard deviations from the mean is considered normal and
non-pathological (van den Berg et al., 2006). Adolescents also differ with respect to the
sequence and tempo of pubertal changes. The duration of puberty varies from one and a
half to six years, and the earlier the onset of puberty, the longer is its duration
(Pantsiotou, Papadimitriou, Douros, Priftis, Nicolaidou, & Fretzayas, 2008; Pinyerd &
Zipf, 2005). The aspect of puberty to which social scientists have paid most attention is
individual differences in pubertal timing.
In addition to the difference in pubertal timing between individuals, there is a difference
between age cohorts. At least in the Western world, average age at puberty has declined
markedly the last century. The age decline seems to have been greatest with regard to the
appearance of pubic hair and breasts (Muir, 2006; Parent et al., 2003). Genetic factors, weight
18
18
5
and body fat, amount of daily exercise, intensive athletic activity, protein intake, illness, and
environmental stress in childhood have all been linked to age at puberty (Burt, McGue,
DeMarte, Krueger, & Iacono, 2006; Graber, Brooks-Gunn, & Warren, 1995). It has been
speculated that better nutrition and lower prevalence of illness has contributed to the so-called
secular trend. These changes have lead to changes in the age at which girls are considered to
experience abnormally early puberty (pubertal praecox or true/central precocious puberty).
1.3 Measuring puberty
Puberty is a process that spans several years and it begins long before there are any overt
signs of it, sometimes earlier than age 6 years (Dorn & Rothenstein, 2004). Puberty is
governed by hormones whose purpose is to generate sexually maturity and to signal this
to others. What is the best way of measuring puberty for research purposes?
Dorn and colleagues (2006) recently published a comprehensive review article on
how to measure puberty. Their answer was that there is no such thing as one measure of
puberty that is always best, but rather that different research questions require different
measures. However, they also argued that it is a limitation of the literature that measures
of puberty differ significantly between studies, not only between disciplines or areas of
research but also within them, even when research questions are nearly the same. This
makes comparisons between studies difficult.
The literature represents many different methodologies concerning puberty. First,
informants, such as girls themselves, parents, and medical staff, have been used.
Different informants, however, do not always agree on where in the pubertal process a
girl is (Schlossberger, Turner, & Irwin, 1992). In most studies, girls themselves or their
parents are used as informants. Self-reports are sometimes criticized for being susceptible
to biases like self-enhancement and self-deception (e.g., Markey, Markey, & Tinsley,
2003). Certain individual characteristics might affect how girls rate their level of
maturity. For instance, obese and non-obese girls rate their breast development
differently (Bonat, Pathomvanich, Keil, Field, & Yanovski, 2002), with obese girls being
more likely to overestimate their breast development than other girls. Medical staff often
use the Tanner ratings procedure, which is used to capture the continuous pubertal
growth process (Marshall & Tanner, 1969). The ratings are based on five stages. Which
of the five stages girls are at is determined by their breast and pubic hair development.
Drawings of girls at different stages are used to see which drawing corresponds most to a
specific girl. During the first Tanner stage, there are no manifest signs of gonadarche or
secondary sexual characteristics. During the final stage, all necessary overt signs of
sexual maturity are present. Menarche typically takes place during one of the later stages.
Having medical staff as informants is costly, and the number of participants in studies
using this methodology is often limited. Therefore, unless the research is clinical,
physical examinations are in many cases not ideal for research purposes. Importantly,
19
19
6
Dorn, Susman, and Ponirakis (2003) showed that different sources of rating, self-reports,
parent reports or physical examinations, lead to different conclusions about the links
between puberty and adjustment. Thus, all informants have limitations, and basing
research on different informants may produce discrepancies in results.
Second, researchers have measured different aspects of puberty, even when having
posed similar research questions. An example of this is the study by Costello et al.
(2007), which showed that overt signs of puberty (i.e., secondary sexual characteristics),
but not hormonal indicators, were linked to youth drinking behavior. Some researchers
have used a single item whereas others have used scales comprising several items, or
aspects, of puberty. Age at menarche is one of the most commonly used measures (Dorn
et al., 2006). Tapping age at menarche is arguably also the easiest way to measure
puberty. It is a discrete event that is easily noticeable and that the majority of girls
accurately recall (Must, Phillips, Naumova, Blum, Harris, Dawson-Huges, et al., 2002).
According to Dorn and colleagues (2006), there are three important factors to consider
when using age at menarche as a measure of puberty in research: First, age at menarche
is a measure of the more advanced phases of pubertal development; thus, it is incorrect to
label those girls who have not yet reached menarche as “prepubertal”. Second, although
there is a tendency for girls to experience menarche late in puberty (i.e. during the fourth
Tanner stage), there are some inter-individual variations; therefore, girls might be at
different levels of maturity concerning secondary sexual characteristics and other aspects
of puberty when they have their first menstrual period. Third, although most girls
remember how old they were when they experienced menarche correctly, this is not
always the case.
The pubertal development scale (PDS), developed by Petersen, Crockett, Richards,
& Boxer (1988), on which adolescents rate their physical development, is the most
common measure of puberty (Dorn et al., 2006). This measure is about perceived
pubertal changes (i.e., breast development, skin, height etc.). Sometimes, girls are asked
whether they feel that they matured, or developed, later, earlier, or at the same time as
their peers. There is some research suggesting that perceived pubertal maturation or
timing might play a greater role in girls’ body image than actual maturation, as measured
by physical examination (Dorn et al., 2003). Finally, endocrine measures have been used,
but not very often in the social sciences. Although puberty is governed by hormones, it is
not until recently that links between hormonal changes and adjustment have been
examined. Contrary to the common belief that puberty is a period of “sturm und drang”,
research suggests that adolescents are not “victims of raging hormones” (Buchanan et al.,
1992). Still, there are some findings that link hormonal levels to aggressive and
depressive affects (e.g. Warren & Brooks-Gunn, 1989). Thus, research on puberty
includes many different measures, and the research question, and also the age period of
20
20
7
interest, should be regarded when deciding upon which aspects of puberty to tap in a
particular study.
Third, different statistical techniques or methodologies have been used. This
concerns the use of different cut-off points. Girls are often divided into early, on-time,
and late maturers. Sometimes, they are divided into two groups, early and late maturers.
Where the cut-off is placed between groups varies markedly between studies. In many
cases, girls are not divided into early, on-time, or late maturers. Instead, pubertal timing
is treated as a continuous scale, and the relation between pubertal timing and other
measures are assumed to be linear. The inconsistency makes comparisons between
studies difficult, which is a possible reason why, in some instances, there are
considerable divergences in findings between and even within studies (Alsaker, 1995).
To avoid misinterpretations, analytic procedures should be carefully described in studies
involving puberty.
Most of the above-mentioned measures tap girls’ pubertal stages. To get a measure
of pubertal timing, researchers typically use one of these measures and control for age.
Either girls are grouped into early, on-time, and late maturers (or just early and late), or
pubertal status is measured on a continuum to reflect whether girls are more or less early
or late. Age at peak height velocity is a measure of pubertal timing that is designed to
reflect how old a girl was when she was growing the fastest. To get a valid measure of
age at peak height velocity, repeated measures at short time intervals (preferably as short
as 6 months should be used. See Dorn et al., 2006)
In sum, social scientists typically measure the overt signs of puberty. But endocrine
aspects of puberty are used at times. When interpreting research findings and
conclusions, it is important to keep in mind which aspect of puberty is under study, and
that the findings and conclusions might only apply under specific circumstances. It is
also important to bear in mind which cut-offs have been used. Different aspects of the
pubertal process might be important for different types of research questions, and some
aspects might play a more prominent role than others. There is not just one way of
measuring puberty. However, striving for concordance between studies with respect to
issues of measurement is important for future development of the various areas in this
research field.
21
21
8
1.4 Adolescent implications of pubertal timing
The following two sections, Adolescent implications of pubertal timing and Adult
implications of pubertal timing, describe the research findings concerning the
implications developing at different time points might have for female development.
Both sections start with somatic implications, with a focus on weight status, and end with
psychosocial implications.
1.4.1 Somatic implications. At what age girls mature seems to have implications for
their weight status. During puberty, girls gain more body fat, but early-developing girls seem
to gain more weight than their peers in adolescence. Early-developing girls are stouter
compared with others (Bini, Celi, Berioli, Bacosi, Stella, Giglio et al., 2000; Biro et al., 2001;
Bratberg, Nilsen, Holmen, & Vatten, 2007a; Villa, Yngve, Poortvliet, Grjibovski, Liiv,
Sjöström et al., 2007), and this difference persists into adulthood (Freedman, Khan, Serdula,
Dietz, Srinivasan, & Berenson, 2003; Garn, LaVelle, Rosenberg, & Hawthorne, 1986;
Hulanicka, Lipowicz, Kozie, & Kowalisko, 2007; Must, Naumova, Phillips, Blum, DawsonHughes, & Rand, 2005; Wang, Zhao, Liu, Recker, & Deng, 2006). It is still unclear whether
the link between developing early and having higher adult BMI (Body mass index) is a result
of early maturation per se, whether it only appears because early-developing girls already
have higher BMI in childhood and because BMI is stable from childhood to adulthood, or
whether it is due to other factors.
High BMI as early as at age 3 has been linked to early puberty (Adair & Gordon-Larsen,
2001; Davison, Susman, & Birch, 2003; He & Karlberg, 2001; Juul, Teilmann, Scheike,
Hertel, Holm, Laursen, et al. 2006; Kaplowitz, Slora, Wasserman, Pedlow, & HermanGiddens, 2001; Lee, Appugliese, Kaciroti, Corwyn, Bradley, & Lumeng, 2007). Furthermore,
most of the research (Campbell, Katzmarzyk, Malina, Rao, Pérusse, & Bouchard, 2001;
Freedman, Khan, Serdula, Dietz, Srinivasan, & Berenson, 2005; Sandhu, Ben-Shlomo, Cole,
Holly, & Davey Smith, 2006) shows that BMI is stable over the lifespan (rs = .40 – .60),
although some studies have failed to obtain these moderate to high correlations over longer
periods of time (Casey, Dwyer, Coleman, & Valadian, 1992). If the link between early
pubertal timing and adult weight status merely reflects the stability of BMI, then it should
disappear once childhood BMI is controlled for. Early research showed a link between
pubertal timing and adult BMI, and assumed that childhood BMI was not the main causal
mechanism (Garn et al., 1986). More recent findings are conflicting. Some studies (Freedman
et al., 2003; Must et al., 2005) controlled for early weight status and found no link, whereas
others have found a link, at least for males (Kindblom, Lorentzon, Norjavaara, Lönn,
Brandberg, Angelhed, et al., 2006).
Given that pubertal timing has an independent impact on adult weight status, a critical
question is whether this depends on other conditions, since not all early-developing women
are sturdily built. A recent Norwegian study (Bratberg et al., 2007a) suggests that since not all
early developers have high postpubertal BMI, there are moderating factors that determine
22
22
9
which early-developing girls will have high postpubertal BMI and which will not. Bratberg
and colleagues (2007a) found that only early-developing girls who had high waist
circumference in early adolescence had elevated levels of BMI in late adolescence. However,
because waist circumference was measured sometime between the ages 12-16, puberty and
waist circumference were confounded in that study. Puberty might already have influenced
waist circumference in some girls at that age, and thus it cannot be thought of as an
independent moderator. Other potential moderators should therefore be explored.
1.4.2 Psychosocial implications. In the 1930s, Jones (1938) and later Jones and
Mussen (1958) started to examine the behavioral implications of pubertal maturation. Those
early studies focused on relations between pubertal timing and personality characteristics in
adolescence and young adulthood. The results indicated that, by late adolescence, earlydeveloping girls had higher levels of social inhibition, shyness and irritability, and a tendency
to have temper tantrums. These results were interpreted as showing that these girls were low
on agreeableness and low on emotional stability. The situation was the opposite for latedeveloping girls. They were high on agreeableness and emotional stability. Since that early
work, the focus has shifted from pubertal implications for personality characteristics to
pubertal implications for adjustment, particularly internalizing and externalizing behaviors
and substance use.
As described in the previous sections, girls differ markedly in terms of the age at which
they experience puberty. As a result, there is a chance that girls are at very different
developmental stages, socially, cognitively and emotionally, when they go through puberty. It
is easy to envisage that being the first in the class to have breasts is very different from having
breasts later than everyone else.
The role of pubertal timing in female adolescent functioning has been the focus of a large
body of literature. Several recent reviews of the literature have concluded that girls’ early
pubertal timing is linked to adjustment problems in a variety of areas (Alsaker, 1995, 1996;
Buchanan et al., 1992; Celio et al., 2006; Connolly et al., 1996; Graber, 2003; Mendle et al.
2007; Stattin & Magnusson, 1990; Susman et al., 2003; Susman & Rogol, 2004). The
associations between pubertal timing and various measures of problematic adjustment seem to
be stronger in early and mid-adolescence and to diminish in late adolescence (Weichold,
Silbereisen, & Schmitt-Rodermund, 2003). Thus, developing early seems to be a risk factor
for female development, at least during the early to mid-adolescent years.
Early-developing girls have more problems in a variety of areas. Internalizing problems,
including depression, social phobia, and eating disorders increase drastically during
adolescence. Girls are affected by these types of problems to a greater degree compared to
boys. During childhood, depression prevalence rates are similar among girls and boys. By
contrast, from early adolescence and throughout life, females are about twice as likely as men
to experience psychological distress, depressive symptoms, and major depressive disorders
(Nolen-Hoeksema, 1991).
23
23
10
Among girls, early maturers seem to be particularly likely to experience internalizing
problems during adolescence. They have more distress, depressive feelings, and more
generalized internalizing difficulties, and are more likely to attempt suicide (Aro & Taipale,
1987; Compian, Gowen, & Hayward, 2004; Ge, Conger, & Elder, 1996, 2001; Graber,
Lewinsohn, Seeley, & Brooks-Gunn, 1997; Kaltiala-Heino, Kosunen, & Rimpelä, 2003;
Patton, Hibbert, Carlin, Shao, Rosier, Caust et al., 1996; Stice, Presnell, & Bearman, 2001;
Wichstrom, 1999; 2000); also, they have a poorer body self-image (Muris, Meesters, van de
Blom, & Mayer, 2005; Williams & Currie, 2000). In one study, early-developing girls were
found to show more depressive symptoms than on-time and late-developing girls every year
from the 7th to the 12th grade (Ge et al., 2001), and at least three studies have found
associations between early puberty and suicidal behavior (Graber et al., 1997; Stattin &
Magnusson, 1990; Wichstrom, 2000). Others have shown that advanced pubertal maturation
at age 11-12 is linked to experiencing stress, but not to other types of psychological
difficulties (Simon, Wardle, Jarvis, Steggles, & Cartwright, 2003). In a recent cross-sectional
study, early pubertal timing predicted higher emotional arousal, which in turn predicted
increased depressive affect (Graber, Brooks-Gunn, & Warren, 2006). At the same time, it
should be said that – when significant associations between pubertal timing and depressive
affect are reported – the effects are low to moderate, and the role of pubertal timing differs
markedly between studies (e.g., Ge et al., 1996; Wichstrom, 1999).
In addition to depressive symptoms, early-developing girls have low body satisfaction.
Body mass is higher on average in early than in late-developing girls (Haynie, 2003), and
early-developing girls tend to be more concerned and dissatisfied with their bodies,
particularly their weight (Cauffman & Steinberg, 1996; Compian et al., 2004; Crockett &
Petersen, 1987; Dorn, Crockett, & Petersen, 1988; Ge, Elder, Regnerus, & Cox, 2001; Graber
et al., 1997; Stattin & Magnusson, 1990; Stice et al., 2001; Zehr, Culbert, Sisk, & Klump,
2007). Some suggest that this might produce eating disorders (Graber et al., 1997; Keel,
Fulkerson, & Leon, 1997). Earlier reviews of the literature considered the association between
poor body image and early pubertal timing to be robustly established (Alsaker, 1995, 1996;
Buchanan et al., 1992). In conclusion, early pubertal timing seems to be associated with
depressive mood and poor body image in adolescence.
The school is an important part of adolescents’ lives. The main purpose of school is
education. However, it also functions as an arena for peer relationship formations, and also
many young people find their first romantic partners at school. Achieving in school is not
only associated with future educational attainments and professional careers; it is also
negatively related to delinquent behavior (Magnusson, Dunér, & Zetterblom, 1975). Earlydeveloping girls have been shown to have more school-related problems, such as nonattendance and lack of school motivation, than other girls, and they are less interested in
academic subjects
24
24
11
(Caspi, 1995; Davies, 1977; Graber et al., 1997; Simmons, Blyth, & McKinney, 1983;
Simmons, Blyth, Van Cleave, & Bush, 1979; Simmons, Carlton-Ford, & Blyth, 1987; Stattin
& Magnusson, 1990). Late maturers, on the other hand, obtain better grades than other girls
during adolescence (Dubas, Graber, & Petersen, 1991). Poor school adjustment among earlydeveloping girls cannot simply be attributed to low IQ, since there are no differences in
cognitive ability between early and late-developing girls (Stattin & Magnusson, 1990). It
might, however, be that involvement in early romantic relationships is one reason why earlydeveloping girls are less well adjusted at school (see below).
Social relationships outside the family become increasingly important during
adolescence. Peer relationships start to play an increasing role early on in adolescence.
Adolescents spend more time with their peers, both in school and during leisure time than
children (see Zimmer-Bembeck, 2002, for a review). Peers are sources of activities, support,
and influence (Savin-Williams & Berndt, 1990) and are usually similar to each other
(Dishion, Patterson, & Griesler, 1994; Hartup, 1996; Kandel, 1978, 1985, 1986). This is due
to the fact that adolescents choose peers who are similar to themselves, a phenomenon that is
referred to as homophiliy. On the other hand, peers also become more similar over time as
they associate with each other.
Research suggests that the peer networks of early-developing girls differ from those of
later maturers. There are studies suggesting that early-developing girls have older peers and
more male peers, and that they associate more with deviant peers, including adult male sex
partners, than do on-time or late-developing girls in mid-adolescence (Costello et al., 2007;
Dick et al., 2000; Ge et al., 1996; Haynie, 2003; Lynne, Graber, Nichols, Brooks-Gunn, &
Botvin, 2007; Mezzich, Giancola, Lu, Parks, Ratica, & Dunn, 1999; Patton, McMorris,
Toumbourou, Hemphill, Donath, & Catalano, 2004; Silbereisen, Petersen, Albrecht, &
Kracke, 1989; Silbereisen & Kracke, 1993; Stattin & Magnusson, 1990; Wichstrom, 2001).
Recent research has shown that high pubertal status among early adolescents strengthens the
association between peer delinquency and the delinquency of the individual (Fergusson,
Vitaro, Wanner, & Brendgen, 2007). Put differently, early-developing adolescents with
deviant peers are more delinquent than late-developing adolescents with deviant peers. This
might be because early-developing girls are more vulnerable to deviant peer influence than
other girls. Other explanations are also possible. For instance, early-developed girls associate
considerably more with chronologically older peer, and these peers are at the peak of juvenile
delinquency, whereas the peers of the later developed girls are chronologically younger and
have not yet reached the age when adolescent problem behavior is highest. Thus, the peers of
early-developed girls might have more problem behavior than the peers of later developed
girls. In sum, early-developing girls seem to be more oriented towards peers, and also to have
peers who are older and therefore more socially advanced (and, perhaps, more problem
behavior).
25
25
12
Another form of social relations that grows in importance, typically starting from early to
mid-adolescence, is relationships with the opposite sex, including romantic relationships. At
the beginning, romantic relationships are typically short-lived and take place in the context of
other peers. As adolescence passes, the nature of romantic relationships becomes more similar
to that of adult romantic relationships (Furman, 2002). By mid-adolescence, most girls and
boys have some experience of dating (Buhrmeister & Furman, 1987; Connolly & Johnson,
1996).
Given the amount of time spent with, talking, and thinking about members of the
opposite sex, it is scarcely surprising that these relationships have important developmental
significance. In some respects, having a romantic partner has been shown to be positive.
Romantic relationships in middle and late adolescence have been associated with lower levels
of social anxiety (La Greca & Harrison, 2005). Early romantic relationships and having many
partners, however, have been associated with an increased risk of problems (ZimmerGembeck, Siebenbruner, & Collins, 2001), which indicates that romantic relationships might
not always be positive for developing youth. Indeed, most studies have shown links between
involvement in romantic relationships and adjustment difficulties (e.g. Darling, Dowdy, Van
Horn, & Caldwell, 1999). Early initiation of romantic relationships has been linked to
depressive symptoms, body image problems, eating disorders, problem behavior, and poor
academic achievements (Cauffman & Steinberg, 1996; Compian et al., 2004; Neemann,
Hubbard, & Masten, 1995; Smolak, Levine, & Gralen, 1993), whereas romantic involvement
in late adolescence has not (Neeman et al., 1995). Thus, early romantic involvement seems to
put girls at risk of internalizing and externalizing problems and substance use.
Although it does not necessarily have to take place in the context of a steady relationship,
sexuality is a key aspect of romantic relationships. Adolescent sexuality is a normative
developmental phenomenon. Most adolescents in western cultures start engaging in sexual
activities, including sexual intercourse, during their teenage years. When in a relationship,
adolescents tend to start by embracing and kissing one another, and then proceed to petting
and more intimate behaviors including intercourse (Miller & Benson 1999).
As with early involvement in romantic relationships, early onset of sexual activity seems
to affect girls negatively. It is associated with having more sexual partners, having intercourse
more often, and having older sexual partners (Moore, Miller, Glei, & Morrison, 1995).
Having sexual intercourse early in adolescence has also been connected with a variety of
problem behaviors (e.g. Pedersen, Samuelsen, & Wichstrom, 2003). Among girls, the
problems range from being depressed and having worse body perceptions to being more
delinquent, having children in adolescence, and drinking more alcohol than same-age, samesex peers (Armour & Haynie, 2007; Harvey & Spigner, 1995; Pedersen et al., 2003). In a
retrospective study, Mott and Haurin (1988) found that once teenagers had started to engage
in sexual activities, they also started to use more substances. Deardorff and colleagues
(Deardorff, Gonzales, & Christopher, 2005) have shown that early sexual activities and early
26
26
13
initiation of substance use are linked to teenage pregnancy. Thus, early sexual activities seem
also to be associated with girls’ problem behaviors.
Puberty is the process of becoming sexually mature, physically speaking. In addition,
girls start to become interested in romantic relationships and sex (e.g. Alsaker, 1996; Furman,
Brown, & Feiring, 1999; Waylen & Wolke, 2004), and feelings of being in love increase
significantly (Larson, Clore, & Wood, 1999). With puberty, girls start to think of and view
themselves differently, and there is an increasing tendency to view one's body as an object for
others to look at and evaluate (Lindberg, Grabe, & Hyde, 2007). When girls think about
physical maturation, they are often aware that their physical development will arouse interest
in boys, and that there are sexual issues connected with their development (O’Sullivan, Heino,
Meyer-Bahlburg, & Watkins, 2000). Accordingly, a large body of research has been devoted
to the connections between puberty and romantic and sexual behaviors.
Early-developing girls move faster into romantic relationships and sexual activities than
other girls. In fact, the link between pubertal status and romantic relationships appears as
early as at age 11 (Compian et al., 2004). Early-developing girls have more experience than
their same-age peers with different aspects of romantic relationships with boys and sexual
issues (feelings of being in love, dating, sexuality, going steady, having older boyfriends,
viewing sexual media, having abortions, and being subject to sexual abuse), and they are more
likely than late developers to attract attention from boys (Simmons, Blyth, & McKinney,
1983). All this appears in studies performed almost six decades ago (Stone & Barker, 1937,
1939) and also, over the past two or three decades, in both North American and European
samples (Aro & Taipale, 1987; Baumeister, Flores, Marín, 1995; Brown, Halpern, & L’Engle,
2005; Cavanagh, 2004; Crockett & Petersen, 1987; Fergusson et al., 2007; Flannery, Rowe, &
Gulley, 1993; Goodson, Evans, & Edmundson, 1997; Haynie, 2003; Jorm, Christensen,
Rodgers, Jacomb, & Easteal, 2004; Phinney, Jensen, Olsen, & Cundick, 1990; RodriguezTomé, Bariaud, Cohen Zardi, Delmas, Jeanvoine, & Slylagyi, 1993; Schor, 1993; Silbereisen
& Kracke, 1993; Simmons & Blyth, 1987; Simmons, Blyth, & McKinney, 1983; Simmons,
Blyth, Van Cleave & Bush, 1979; Smith, Udry, & Morris, 1985; Stattin & Magnusson, 1990;
Turner, Runtz, & Galambos, 1999; Udry, 1979, 1988; Udry & Billy, 1987; Zelnik, Kantner, &
Ford, 1981). It also appears in non-western cultures, such as Hong Kong (Lam, Shi, Ho,
Stewart, & Fan, 2002) and North Sudan (Otor & Pandey, 1999). Thus, the link between
pubertal development and initiation of sexual relationships seems to be quite robust, and it
seems well-established that girls who mature early enter into romantic and sexual
relationships earlier than their same-age peers. Against the backdrop of the negative
developmental consequences of early sexual behavior and early romantic involvement, it
seems reasonable to assume that early-developing girls are therefore over-represented among
girls in problematic adjustment situations.
During no other period in life do girls engage in delinquency to the extent that they do
during adolescence. In fact, minor delinquency seems to be part of normative adolescence,
27
27
14
with the vast majority breaking the law by drinking or committing petty vandalism (Rutter &
Giller, 1984). In one study, 80% of the teenagers between 11.5 and 15 years were found to
have engaged in problem behaviors during the preceding month (Maggs, Almeida, &
Galambos, 1995). Problem behavior peaks around age 16-17, and then declines in prevalence
(Sampson & Laub, 2003).
In accordance with what was presumed above, a large body of literature has established
that early-developing girls are over-represented among girls with delinquency, substance use,
and other forms of problematic conduct in adolescence. Well over a hundred empirical studies
have been devoted to this issue the last 20 years, some of which have made significant
contributions to the understanding of the role played by pubertal timing in adolescent girls’
problem behaviors (e.g., Aro & Taipale, 1987; Burt, McGue, DeMarte, Krueger, & Iacono,
2006; Caspi, 1995; Caspi, Lynam, Moffitt, & Silva, 1993; Caspi & Moffitt, 1991; Dick, Rose,
Viken, & Kaprio, 2000; Ge, Brody, Conger, & Simons, 2006; Ge, Brody, Conger, Simons, &
Murry, 2002; Ge, Conger, & Elder, 1996; Ge, Jin, Natsuaki, Gibbons, Brody, Cutrona, et al.,
2006; Graber, Lewinsohn, Seeley, & Brooks-Gunn, 1997; Graber, Seeley, Brooks-Gunn, &
Lewinsohn, 2004; Haynie, 2003; Lanza & Collins, 2002; Lynne, Graber, Nichols, BrooksGunn, & Botvin, 2007; Martin, Kelly, Rayens, Brogli, Brenzel, Smith et al., 2002; McMaster,
Connolly, Pepler, & Craig, 2002; Obeidallah, Brennan, Brooks-Gunn, & Earls, 2004; Patton,
McMorris, Toumbourou, Hemphill, Donath, & Catalano, 2004; Silbereisen & Kracke, 1993;
Silbereisen & Kracke, 1997; Simmons & Blyth, 1987; Simmons, Blyth, & McKinney, 1983;
Simmons, Blyth, Van Cleave, & Bush, 1979; Simmons, Carlton-Ford, & Blyth, 1987; Stattin
& Magnusson, 1990; Tschann, Adler, Irwin, Millstein, Turner, & Kegeles, 1994; Wichstrom,
2001; Wiesner & Ittel, 2002; Wilson, Killen, Hayward, Robinson, Hammer, Kraemer et al.,
1994). As early as at age 11, early pubertal timing is linked to risky and health-compromising
behaviors (e.g. smoking, drinking, and sexual behaviors; Markey et al., 2003). Moreover,
early-developing girls have higher rates of experimental and regular tobacco use,
externalizing problems, norm violations, delinquency, drug and alcohol use, sensation
seeking, antisocial personality, and psychopathology, aggression, school problems, including
non-attendance and lack of school motivation, engagement with deviant peers, sexually
harassing other-sex peers, and peer support for problem behavior (see Alsaker, 1995, 1996;
Buchanan et al., 1992; Celio, Karnik, & Steiner, 2006; Mendle, Turkheimer, & Emery, 2007;
Susman et al., 2003; Susman & Rogol, 2004; Pinyerd & Zehr, 2005 for reviews). Different
behavioral pathologies, such as conduct disorders, are also linked to early puberty (Burt et al.,
2006).
In addition to the steadily growing literature on the connection between pubertal timing
and various types of problematic conduct on behalf of early-developing girls, recent findings
suggest that there is also a connection between pubertal timing and physical and violent
victimization (Haynie & Piquero, 2006). Controlling for girls’ and boys’ own violent
delinquency, early maturers of both genders reported higher levels of victimization than other
28
28
15
girls and boys. Early-developing girls are also more exposed to maltreatment in childhood
than other girls (Costello et al., 2007). These studies show that it is not only that earlydeveloping girls engage in problem behaviors; they are more exposed to them as well.
Although puberty has psychological and social meaning, it only refers to biological
maturation and does not necessarily entail maturation in other areas. For instance, puberty
might occur long before the girl herself acknowledges the new phase she is entering, in the
sense that she also feels or sees herself as mature. As a group, however, early-developing girls
feel more mature than their later-developing peers (Stattin & Magnusson, 1990).
One plausible reason for why pubertal timing is more important for girls’ adjustment than
pubertal status is the importance of social comparisons in adolescence (Alsaker, 1995). This
suggests that girls’ cognitions and perceptions about puberty and their own maturation
relative to others may play important roles in the link between pubertal timing and
adjustment. The idea is that early developing girls may be viewed and view themselves as
different from their peers, which might have negative consequences for them (Siegel, Yancey,
Aneshensel & Schuler, 1999). In fact, scholars have argued that it is when adolescents feel
more mature than their same-age peers that pubertal timing plays a meaningful role in their
psychosocial adjustment (e.g., Alsaker, 1995). This idea is in line with recent research
showing that feeling older than one’s age and subjective maturity is linked to antisocial
behavior, substance use, and sexual behavior (Arbeau, Galambos, & Jansson, 2007) and a
study in which we showed that subjective maturity is linked to antisocial behavior
(Andershed, Johansson, & Pepler, 2006). Thus, subjective feelings of maturity might be an
important factor in the link between pubertal timing and adjustment.
As maturity has many different meanings to adolescents, like responsible behavior,
political views, religious activity, physical maturity, power, and socially advanced behaviors,
such as drinking, having serious romantic partners or dressing sexy (Tilton-Weaver, Vitunski,
& Galambos, 2001), it is possible that there are subgroups of girls who feel mature for
different reasons, and that the adjustment situations of these subgroups might also be
different. Assuming that feeling mature is negative might be overly simplistic. Instead, since
the reasons for feeling mature may differ between individuals, feeling mature might have
different implications. More research is needed to understand if this is the case.
Researchers have studied girls’ pubertal timing in relation to adolescent adjustment in
many different areas. Early developing girls have been examined as both agents of action and
recipients of actions. The major reviews in this field have all concluded that evidence
demonstrating that early pubertal timing is related to adjustment difficulties in several
domains has amassed over the last couple of decades (Alsaker, 1995, 1996; Buchanan et al.,
1992; Celio et al., 2006; Connolly et al., 1996; Graber, 2003; Mendle et al. 2007; Stattin &
Magnusson, 1990; Susman et al., 2003; Susman & Rogol, 2004). The reasons for this remain
unclear. There are, however, reasons to believe that early romantic and sexual involvement,
and also subjective feelings of maturity, might be important pieces in the puzzle.
29
29
16
1.5 Adult implications of pubertal timing
1.5.1 Somatic implications. As noted in the section on the implications of pubertal
timing during the adolescent years, pubertal timing appears to have somatic implications.
Early-developing women are at higher risk of certain diseases, such as breast cancer and
cardiovascular disease. Experiencing menarche before age 12 increases the risk of breast
cancer by 50% compared with experiencing menarche at age 16 (Grumbach & Styne, 2003).
In adolescence, early-developing girls are stouter than other girls (Adair & GordonLarsen; Bini et al., 2000; Biro et al., 2001; Bratberg, Nilsen, Holmen, & Vatten, 2007b; Villa
et al., 2007), and this difference persists into adulthood. In several studies, women who were
the first in their cohort of peers to experience puberty had a higher adult BMI than laterdeveloping women (Freedman et al., 2003; Garn et al., 1986; Hulanicka et al., 2007; Must et
al., 2005; Wang et al., 2006). Whether this simply depends on childhood BMI or not, is still
controversial. Similarly, little is known about factors that might interact in the link between
pubertal timing and adult weight status.
1.5.2 Psychosocial implications. Some of the effects of early puberty on
psychological and social factors seem to be limited to adolescence whereas others seem to
linger on into adulthood (e.g., Celio et al., 2006; Graber et al., 2004; Stattin & Magnusson,
1990). There also findings that indicate that the situation in adulthood might be reversed
compared with that in adolescence. In a study by Dubas (2003), early-developing girls
manifested the most delinquency, and late-developing girls the most depression in
adolescence. By contrast, in young adulthood (sometime between the ages 19 and 26), earlydeveloping women manifested the most depression, and late-developing girls the most
delinquency. The important contribution of this study is that it shows that the role of pubertal
timing might be very different in adolescence and adulthood, and it sheds light on the
importance of long-term follow-up studies of the role of pubertal timing in female
development.
According to Elder (1994), salient events during critical periods in people’s lives have the
potential to shape life-course trajectories. I noted earlier in the Introduction that puberty in
many ways is a salient event that takes place during a critical period in the lives of adolescent
girls. To the extent that prior research linking pubertal timing with adjustment difficulties is
correct, based on Elder’s view, we should expect pubertal timing to shape female
developmental trajectories into adulthood. But, is this true?
The simple answer is that we do not know. While the literature on the links between
pubertal timing and its developmental significance in adolescence is substantial, the opposite
is true for the literature on the links between pubertal timing and its developmental
significance in adulthood. There is no satisfactory answer to the question of the role played by
pubertal timing in psychosocial adjustment in adulthood. There are some studies that have
tried to answer this question, but often the follow-ups have only been for a couple of years at
best, and the existing findings are conflicting.
30
30
17
There are two main ways of thinking about the long-term consequences of pubertal
timing. One is that the effects of pubertal timing are limited to the period when early or late
developers are developmentally deviant, which is during early and mid-adolescence. This
view is supported by some research showing that pubertal timing is linked to school
performance in early to mid-adolescence, but not in late adolescence (Dubas et al., 1991). The
other way of thinking is that pubertal timing will continue to play a role in adult life. This
view is supported by research showing that early-developing girls have children earlier than
other girls (Stattin & Magnusson, 1990). Furthermore, pubertal timing alone might not predict
adult adjustment. Perhaps it is only when early puberty co-occurs with certain other factors,
like early romantic involvement, that it is predictive of adjustment in adulthood.
Prior research that has followed females over long periods of time has focused on
psychopathology, delinquency, family, and career development (e.g., Dubas, 2003; Graber et
al., 2004; Stattin & Magnusson, 1990). Some research has shown that early and latedeveloping women in young adulthood seem to be equally likely to cohabitate or be married
(Graber et al., 2004; Stattin & Magnusson, 1990). Other research, however, has shown that
early maturers are more likely to marry at an early age than other women (Sandler, Wilcox, &
Horney, 1984). A large body of research on the timing of childbearing has shown that girls
with early menarche are more likely to become pregnant and to have children early in life,
including the teenage years, than girls who develop later (e.g., Deardorff et al., 2005;
Hockaday, Crase, Shelley, & Stockdale, 2000; Manlove, 1997; Stattin & Magnusson, 1990;
Woodward, Fergusson, & Horwood, 2001). This is not to say that all studies have shown this
association. For instance, Graber and colleagues (2004) did not report any differences in
motherhood at age 24 between women whose puberty was early or late. Thus, although the
majority of studies have shown that early-developing women have children earlier but do not
marry earlier, others have failed to do so.
Research findings concerning early developers’ educational attainments are also
inconsistent. No difference has been reported between early and on-time developers in terms
of educational level at the 12th grade or at age 24 (Dubas et al., 1991; Graber et al., 2004). By
contrast, one study (Stattin & Magnusson, 1990) found that, as young adults, early developers
had lower educational accomplishments compared with other women. In addition, they had
work positions that required lower educational levels than did other women. This could not be
explained by lower intelligence or level of parental education.
Finally, researchers have failed to find more social adjustment problems among earlydeveloping women in young adulthood (Dubas, 2003; Graber et al., 2004). In one study,
however, early-developing women were over-represented in official criminal records between
the ages 18 to 33 years (Stattin & Magnusson, 1990). In most areas of adjustment that have
been examined, the findings are conflicting. Thus, there is no straightforward answer to the
question of what adulthood holds for early-developing girls, and further longitudinal research
is clearly needed.
31
31
18
1.6 Explaining the link between pubertal timing and adjustment
The reasons why early-developing females are at greater risk of adjustment difficulties are
still largely unknown. Researchers have forwarded different hypotheses to explain how
pubertal experience is related to adjustment. These hypotheses differ extensively on the
assumptions they make about development, whether it is puberty status or pubertal timing that
is assumed to be linked to adjustment, and which girls (early, on-time, or late maturers) are at
greatest risk of having problems.
Explanations can be grouped into three main categories: biological, psychosocial, and
selection (Mendle et al., 2007). Biological theories focus on the dramatic hormonal changes
during the early stages of puberty, since these are thought to have a direct influence on
behavior. Psychosocial explanations are the most commonly used in the social sciences. They
focus mainly on contextual influences. A basic idea is that early maturation brings about
social changes (e.g., different expectations from adults and different peer networks) for which
girls are not cognitively, psychologically, and socially prepared. Some also argue that
deviating from the norm, whether it is developing earlier or later than peers, is in itself linked
to difficulties.
The selection explanation is based on the fact that early maturation is largely due to
genetic factors. Thus, an early-developing mother is likely to have an early-developing
daughter. Because early maturers are more likely to have adjustment problems, it is difficult
to disentangle the effects that early maturation might have from genetic and family
environmental factors. Research, however, suggests that, at least for smoking and alcohol
drinking, genetic and family factors do not explain the link between early development and
problem behaviors (Dick et al., 2000). Similar conclusions, that there does not seem to exist a
common genetic influence on pubertal timing and problem behavior, have been reported by
Burt et al. (2006) and Jorm et al. (2004). Moreover, it is thought that the social consequences
of early puberty are more important in explaining the link between pubertal timing and
adjustment problems than physiological factors (Steingraber, 2007).
The most common hypotheses that have been forwarded concerning the link between
pubertal timing and psychosocial adjustment problems in the social science literature are the
off time hypothesis, the early timing hypothesis, and the accentuation hypothesis (Caspi &
Moffitt, 1991; Dubas, 2003; Ge et al., 1996). According to the off time hypothesis (Eichorn,
1975), or the maturational deviance hypothesis (Alsaker, 1995; Weisner & Ittel, 2002),
experiencing puberty earlier or later than most peers is stressful, and will lead to subsequent
adjustment difficulties. On the other hand, developing at the same time as most others should
not be linked to any particular problems.
The early timing hypothesis is similar to the off time hypothesis. It predicts that early
puberty is linked to problem behavior. A variation of the early timing hypothesis is the stage
termination hypothesis (Peskin, 1973; Weisner & Ittel, 2002). The idea here is that when
puberty starts early, psychological development is interrupted so that girls are less prepared to
32
32
19
resolve the developmental tasks of adolescence. The girls are too emotionally immature and
have too limited experience to successfully face the new challenges. According to Peskin
(1973), the asexual latency period, during which ego mechanisms needed for coping with the
sexual drive of puberty are developed, is substantially shortened for early developing girls. In
addition, early-developing girls’ appearances might elicit expectations from adults that are
excessively based on the girls’ cognitive, emotional, and social skills. This discrepancy might
be stressful for the developing girl. Both the off-time and early timing hypotheses suggest that
being out of synch with peers creates internal distress, which in turn spurs girls to engage in
problem behavior. Much emphasis is placed on the emotional stress that comes with
developing out of sync with one’s peers (Brooks-Gunn & Graber, 1994; Ge et al., 2001;
Steinberg & Morris, 2000). Research has failed, however, to show that distress explains the
link between early puberty and problem behavior (Tschann et al., 1994; Weisner & Ittel,
2002). Thus, there is still little empirical support for the early timing and the off time
explanations.
According to the accentuation hypothesis, periods of discontinuity will accentuate
differences between people (Caspi & Moffitt, 1991). Thus, in times of stress and uncertainty,
people’s behavioral dispositions will manifest themselves in behavior more clearly than
during normal circumstances. Consequently, puberty should not have the same effect on all
girls, but should enlarge individual differences that existed before puberty. In the study by
Caspi and Moffitt (1991), early-developing girls were found to increase with respect to
problem behavior more than on-time and late-developing girls, but this increase was greatest
for the early-developing girls with high problem levels before puberty. The accentuation
hypothesis does not explain why early-developing girls adopt problem behaviors more than
late-developing girls, but it suggests that the acceleration in problem behavior is dependent on
conditions that were already at hand much earlier. Nor has the hypothesis, to my knowledge,
been supported in other research. It is important to note that none of the hypotheses
mentioned make predictions about whether the effects of pubertal timing will be limited to
adolescence or whether they will continue into adulthood.
33
33
20
1.7 Remaining questions – limitations of the existing literature
Despite the large body of literature on the physiological, psychological and social meanings
of female pubertal timing, many questions remain to be answered with respect to the
implications of female pubertal timing. There are three major limitations to the literature. One
is a lack of good explanations for the link between early puberty and problem behavior. The
second is a lack of comprehensive research that includes biological, psychological and social
factors, and their interactions, when predicting any given outcome. The third is a lack of longterm follow-ups and ideas about the longer-term implications of being an early or laterdeveloping girl.
A lack of satisfying explanations and direct tests of the mechanisms that explain the wellestablished link between pubertal timing and adjustment is one major limitation of the female
pubertal timing literature. The current literature provides few comprehensive explanations of
the mechanisms linking pubertal timing and concurrent and subsequent adjustment. As far as
early pubertal maturation puts adolescents at a higher risk of adjustment difficulties, as has
been shown repeatedly in previous studies (e.g., Alsaker, 1995, 1996; Buchanan et al., 1992;
Celio, Karnik, & Steiner, 2006; Graber, 2003; Mendle et al. 2007; Stattin & Magnusson,
1990; Susman et al., 2003; Susman & Rogol, 2004; Waylen & Wolke, 2004 for reviews),
understanding of why and how early development is linked to difficulties, is crucial. The
hypotheses that have been forwarded do not give adequate explanations of why earlydeveloping girls, as opposed to other girls, display more adjustment difficulties in general and
in particular with respect to social processes. The explanations have rarely been tested and are
often used for explaining findings post-hoc than predicting them.
A limitation related to the lack of established mechanisms is the paucity of research with
a biopsychosocial approach to studying development. In the literature, pubertal timing has
often been studied as the sole predictor, not as one of many interacting factors or as one part
of a whole. Most of the research in the puberty area has been based on a main effects
approach (see Graber, 2003). However, there are few reasons to believe that direct effects
models are the most appropriate for studying the implications of pubertal timing. Rather,
interactive models, including psychological and social factors, should be used. All evidence
and all logic point to this. This is highlighted by Graber (2003) who stated that “Interactive
models that account for either internal characteristics of the individual or variations in the
individual’s environment are the only approaches to explaining puberty-psychopathology
links that have any basis in logic… ” (p. 309). Still, the majority of studies have been based
on the main effects approach.
The final major limitation in the female pubertal timing literature is a lack of long-term
follow-ups. There is little knowledge of the long-term implications of female pubertal timing.
Very few studies have examined the role that pubertal timing has for social adjustment in
adulthood, and even fewer have utilized prospective longitudinal designs. The problem this
entails is that there is little information on whether the potential effects of pubertal timing in
34
34
21
adolescence linger on into adulthood or whether they are adolescent-limited. Nor is there
much information on whether early and late-developing girls show equal levels of problem
behavior over time. This issue was addressed by Dubas (2003), who investigated associations
between pubertal maturation and adjustment in a Dutch sample that was followed from
adolescence into young adulthood. In her study, early-developing girls were found to manifest
the most delinquency in adolescence. By contrast, in young adulthood (sometime between the
ages 19 and 26), it was the late-developing girls who manifested the most delinquency. Thus,
it is possible that late-developing girls display higher levels of problem behavior in late
adolescence and/or adulthood, rather than in early and mid-adolescence. With the design used
by the vast majority of studies, it is not possible to get information about whether or not this is
the case.
Many questions remain as to what future implications girls’ pubertal timing will have.
There are several possible answers to these questions. Either pubertal timing only matters in
adolescence and thereafter is of little importance, or it matters in adulthood too. If the first is
true, this would imply that later-developing females will catch up with their earlierdeveloping counterparts in terms of problem behavior, or that the early-developing girls
would adjust to the behavioral norm (i.e., display similar behavior as most others). If the
second is true, then the differences between early and late-developing females would remain
in adulthood. There is, however, a third possibility. It may be that pubertal timing has longterm implications, but only under certain circumstances. This third possibility concerns
temporal conditions during which the implications of pubertal timing differ. For instance,
research on the adolescent years has shown that early pubertal timing is linked to aggressive
behavior only if girls live in disadvantaged neighborhoods (Obeidallah et al., 2004). Thus,
early puberty might have long-term implications for certain subgroups of girls, but not for
others.
In sum, the key remaining research questions, which concern soma et psyche, in the
female pubertal timing literature are:
•
Which mechanisms explain the connections between pubertal timing and
concurrent and subsequent adjustment?
•
Is the impact of pubertal timing different under different contextual conditions?
•
What are the long-term effects of pubertal timing?
35
35
22
2. Aims of the dissertation
This dissertation was designed to extend the prior literatures on female pubertal timing and its
concurrent and future implications. Its aim is to investigate the short and long-term
developmental significance, concerning soma et psyche, of female pubertal timing. It has
three purposes. One is to examine mechanisms that might explain the links between pubertal
timing and concurrent and future adjustment. Another is to use biopsychosocial models to
explore the impact of pubertal timing on behavior. The third is to study the long-term
implications of girls’ pubertal timing, with respect to soma et psyche.
2.1 The biopsychosocial approach
In the literature, pubertal timing has often been studied as a sole predictor, not as one of
several interacting factors or as part of a larger picture. People, however, are biological,
psychological, and social beings, and adolescence is a stage in biopsychosocial development.
Thus, biological, psychological, and social factors should all play important roles in shaping
developmental trajectories. In order to get a better understanding of girls’ adolescent
development, we need to consider their functioning and development in its entirety.
The core of developmental psychology lies in person-environment interactions at
different levels over time (Magnusson, 1988; Magnusson & Stattin, 1998). Studies in the
pubertal timing literature, using direct effects models, have examined direct associations
between pubertal timing and given outcomes, such as externalizing behavior and substance
use. However, puberty is undoubtedly a biological event that takes place in a social context.
Therefore, it seems reasonable to expect that pubertal timing will have different implications
in different social contexts, and perhaps also in different psychological contexts (i.e., with
regard to individual differences). The biopsychosocial model of development (e.g.
Bronfenbrenner & Ceci, 1994; Cairns, 1979; Ford & Lerner, 1992; Magnusson, 1988;
Magnusson & Stattin, 1998) forms the theoretical foundation of this dissertation. According
to this model, biology, psychological factors, and social context are all significant contributors
to an individual’s functioning and development. This is a holistic model of development. It
requires that biological, psychological and social factors are considered together to obtain a
comprehensive understanding of individual development and functioning, and it entails that it
is erroneous to neglect any one of these domains. The idea is that interaction between these
three domains of functioning clarifies important developmental processes. It follows from this
idea that the effects of physiological systems are not autonomous and isolated, but are
responsive to psychological and social factors. Thus, when adopting a biopsychosocial
approach, rather than being very general, researchers strive to make statements about
development and behavior that are specific and concern subgroups of individuals in their
developmental contexts.
36
37
23
Applied to the study of psychosocial implications of pubertal maturation, the model
requires that puberty should be examined together with psychological factors and the social
context that developing adolescents are in for the achievement of a better understanding of its
developmental significance. One could also argue that the model suggests that biological, or
physical, disposition creates an inclination for adolescents to behave in certain ways and that
this behavioral tendency is manifested in certain social contexts. Thus, person-environment
interactions should be key research objects in research on pubertal maturation.
According to the biopsychosocial model of development, (pubertal) timing is not
everything. A thorough understanding of the implications of female puberty requires
consideration of psychological and social factors that act alongside maturational timing.
Neglecting to do so might conceal the possibility that the link between pubertal timing and
any given somatic or psychosocial outcome depends on other conditions (biological,
psychological and/or social), and that early-developing females’ concurrent and future lives
might be radically different because of them.
2.2 The peer socialization hypothesis
A limitation that is related to the paucity of research in which a biopsychosocial approach to
the study of development has been adopted is a lack of established mechanisms through
which pubertal timing is related to a behavioral outcome. The peer socialization hypothesis
was advanced by Stattin and Magnusson (1990) to offer such a mechanism and it states that
early-developing girls will tend to affiliate with peers who are similar to them in biological
maturation, and that these will be chronologically older peers. In addition, early-developing
girls will establish romantic relationships with boys earlier than their same-age peers, most
often with older boys, because they perceive older boys as similar to them in maturity (Rowe
& Rodgers, 1994). Because older adolescents are on average engaged in more problem
behavior than younger adolescents (Sampson & Laub, 2003) and boys have more problem
behavior than girls, girls who join older peer groups and who have romantic relationships with
boys at an early age will have peers who have more problem behavior than their same-age
peers, and these older peers and boyfriends will socialize them into the same kinds of
behaviors. The girls will be encouraged and supported in this kind of behavior by their peers,
and will to a greater extent be brought into leisure-time settings in which these types of
behaviors are normative. Thus, affiliating with advanced peers, including older boyfriends, is
a mechanism through which early developing girls develop problem behavior. Although the
peer socialization hypothesis takes into account relationships with peers of both genders,
romantic relationships with boys should play a more important explanatory role with respect
to problem behavior. These are potentially the most negative influences, because on average
boys score higher than girls on problem behavior at every age. Note that the peer socialization
hypothesis places no emphasis on internal stress, as the earlier mentioned hypotheses do, but
focuses instead on the implications that early puberty has for peer relationships (Magnusson,
38
37
24
Stattin, & Allen, 1985, 1986; Stattin & Magnusson, 1990; Caspi et al., 1993). In sum, the peer
socialization hypothesis states that the link between early development and problem behavior
appears because early-developing girls seek out similar peers in the same way that everyone
else does, but in so doing, they find themselves in older, more socially advanced peer groups
and they adopt the behavior which is normative in these contexts.
The peer socialization hypothesis has the advantage, compared with other hypotheses, of
being based on established principles of peer affiliation and influence, and previous
longitudinal research supports this hypothesis. In one study, the early-developing girls’
problem behavior was found to be largely due to greater romantic interests and relationships
(Stattin & Magnusson, 1990). In another study, controlling for relevant selection factors, the
highest self-reported delinquency and number of delinquent peers were found among the
early-developing girls who had entered mixed-sex schools, where there is ample opportunity
for meeting boys, rather than among all-girls schools (Caspi, 1995). In a third study, having
delinquent peers at Grade 6 was found to mediate the association between perceived pubertal
timing and delinquency and aggression at Grades 6 through 8 (Lynne et al., 2007). In a
Norweigan study (Wichstrom, 1999), pubertal timing was linked to frequent alcohol
intoxication, a link that was mediated by peer problem behavior. Haynie (2003) found that
time spent with peers, involvement in romantic relationships, and having deviant peers,
together with parental relations, mediated the link between pubertal timing and party
delinquency. Finally, early developing girls who had substance-using peers were more willing
and had a greater intention to use substances themselves two years later than later-developing
girls with substance-using peers (Ge et al., 2006). Although the longitudinal studies that have
examined the role of peers in the link between early puberty and problem behavior have had
different focuses, these findings are all in line with the peer socialization hypothesis.
There are, however, some basic issues that are still in question. One of the most critical is
whether romantic relationships with boys play a mediating role, or are themselves problem
behaviors that have some other cause. Another is whether independent reports of peers’
behaviors will yield the same results as past studies, where information about peers has come
from the girls themselves. A third is whether the older peers who enter into relationships with
younger girls are more problematic than others of their age. A fourth remaining question
concerns where early-developing girls meet the peers that eventually might socialize them
into problem behavior. These unanswered questions indicate that the peer socialization
hypothesis still needs more empirical support to be accepted as an explanation for the
differences in social adjustment between early and later developed girls. Perhaps the most
critical remaining question concerns the contextual conditions under which we would expect
the processes described by the peer socialization hypothesis.
38
39
25
2.2.1 The role of contexts. One idea that has been forwarded is that the implications
that early puberty has for behavior depend on contextual conditions. The contextual
amplification hypothesis is a person-environment interactional hypothesis, which deals with
the link between early puberty and problem behavior (Ge, Natsuaki, Jin, & Biehl, 2007). It
has been developed against the backdrop that associations between early puberty and problem
behavior have not always been found, and – even when found – the magnitude of the effect of
pubertal timing on problem behavior is modest. Ge and colleagues (2007) argue that the
explanation for these findings is that the role of pubertal timing in development depends upon
the social contexts in which puberty occurs. In other words, contextual conditions are thought
to play a central role in moderating the effects of pubertal timing on girls’ problem behavior.
It is assumed that early puberty is particularly detrimental when the surrounding context is
adverse (i.e. disadvantaged neighborhood, harsh parenting etc). By contrast, the chances that
early-developing girls will have problems decrease if the environment is supportive. Thus,
according to this hypothesis, the link between pubertal timing and problem behavior will be
more or less strong depending on certain features of the context. In a poor social context, the
link between early puberty and problem behavior should be amplified. In line with this
general idea, research has shown that geographic residency (rural versus urban area),
neighborhood, (advantaged versus disadvantaged, where girls are more likely to encounter
deviant peers), and type of school (girls only versus mixed, where there are more
opportunities to meet boys) all seem to moderate the link between early puberty and problem
behavior (Caspi et al., 1993; Dick et al., 2000; Obeidallah, Brennan, Brokks-Gunn, & Earls,
2004). Neighborhood also seems to moderate the link between early puberty and associating
with deviant peers (Ge et al., 2002). In fact, drs have frequently been referred to as an
important feature of a context that increases the negative impact of early puberty in this
literature (Caspi et al., 1993; Ge et al., 2002; Obeidallah et al., 2004). Thus, contextual
circumstances might increase or decrease the risks that are associated with early puberty.
Precisely why early-developing girls are more negatively affected by poor contexts than
other girls is not specified in the contextual amplification approach. This approach considers
the conditions under which the early puberty-problem behavior link is most likely to occur,
whereas the peer socialization hypothesis considers the mechanisms linking early puberty and
problem behavior. Thus, the peer socialization and contextual amplification hypotheses could
be viewed as complementary, and – if verified – provide a better understanding of the link
between early puberty and adjustment difficulties among girls than is offered by other
theoretical models.
40
39
26
2.3 Purpose of the papers in the dissertation
The main aim of this dissertation is to investigate the short and long-term developmental
significance, concerning soma et psyche, of female pubertal timing. It has three purposes. The
first is to examine mechanisms that might explain the links between pubertal timing and
concurrent and future adjustment. This is the aim of Paper I. In that paper, we report on the
development and testing of the peer socialization hypothesis in four different studies using
five different samples. The main idea of the peer socialization approach is that the reason why
early-developing girls have been found to have relatively high levels of problematic conduct
in adolescence is that they start engaging with older peers and boyfriends, and – through them
– are socialized into more socially advanced behaviors, including problem behavior. This
hypothesis builds on previous research showing that people seek out others who are like them,
and that romantic relationships in adolescence are linked to problem behavior. It is
hypothesized that these links are strongest in contexts where there is an over-representation of
deviant youth and in cultures where heterosexual relationships among adolescents are
sanctioned.
The second purpose is to use a biopsychosocial model and a person-oriented approach to
explore the developmental significance of pubertal timing. This is done in Paper II. The idea
underlying the paper is that more accurate information would be obtained if biological,
psychological, and social predictors are considered together, rather than if pubertal timing is
considered as a sole independent variable. In Paper II, I examine a biological factor (girls’
ages at menarche), a psychological factor (girls’ self-perceptions of maturity), and a social
factor (early romantic relationships with boys) simultaneously in an attempt to obtain a more
comprehensive understanding of the role of pubertal timing in girls’ adjustment. In Paper I,
we assume that early-developing girls feel more mature than their same-sex peers and that
they therefore affiliate with older peers. For the study reported in Paper II, girls’ perceptions
of their own maturity are included in the analyses. It is hypothesized that early-developing
girls who feel mature and who have early romantic relationships with boys will have the most
problematic adjustment. Scholars have suggested that the reason why pubertal timing is linked
to adjustment is that early-developing youth feel different from other youth (e.g., Alsaker,
1995). Maturity can mean many different things to young people, however (Tilton-Weaver et
al., 2001). Socially advanced behavior is certainly one, but responsible behavior is another.
Therefore, I argue that early-developing girls do not act more deviantly simply because they
feel mature. They might just as well act maturely and responsibly. Thus, it is hypothesized
that early-developing girls who do not have early sexual relationships, but who feel mature,
should not have particularly high levels of problematic conduct.
Because there is a dearth of long-term follow-ups of females with early and late puberty,
the third purpose of this dissertation is to study the long-term implications of girls’ pubertal
timing. Findings are presented in papers III and IV, based on two different prospective
longitudinal samples. The implications of interest concern soma et psyche. In Paper III, we
40
41
27
ask what future implications pubertal timing has for females’ weight status, body perception,
and quality of life. Against the backdrop of the literature showing that early-developing girls
are more depressed in adolescence, it is of great importance to understand whether early
maturers will have lower subjective well-being and perceived poorer quality of life in
adulthood. To my knowledge, no study has followed females from middle childhood to
midlife in order to examine the role played by pubertal timing with respect to midlife life
satisfaction. In Paper IV, we pose four different questions. The two main questions are
whether pubertal timing has a unique predictive impact on adult weight status after controlling
for childhood weight status, and whether appetite moderates the link between pubertal timing
and adult weight status. The first question is posed because of the controversy concerning the
role of childhood weight status in the link between pubertal timing and subsequent weight
status. The second question is posed because previous research has shown that it is likely that
there are meaningful subgroups of early-developing girls whose adult weight statuses might
be radically different. Appetite has previously been linked to later body mass, even after
controlling for concurrent body mass (Lee & Song, 2007). Thus, if early pubertal timing
predicts adult weight status, there are reasons to believe that a large appetite may strengthen
the link between early pubertal timing and adult weight status. Because of the nature of Paper
IV (long-term follow-up and integration of biological and psychological factors), it meets
both the second and the third purposes of this dissertation.
42
41
28
3. Method
3.1 Participants and procedures
This dissertation uses six different samples. Paper I uses five different samples and Papers II,
III, and IV use one sample each. The samples differ in terms of: (1) size (Ns range from 90 to
955), (2) the age of the participants at time of the data collections (age ranges from 8 to 43
years), (3) when the information was collected (the earliest information used in this
dissertation was collected in the 1960th and the latest data was gathered in 2002), and (4)
where the data was collected (in three different municipalities in Sweden and in one
municipality in Slovakia). Taken together, the samples capture girls in different locations, at
different time periods, and at different ages. Given that the findings will point in the same
direction, the chances that they depend on sample characteristics are small. Table 1 shows an
overall description of the samples.
3.1.1 Sample I. Sample I was used for papers I, II, and III. This sample comes from the
longitudinal research program Individual Development and Adaptation (IDA). The IDA
project was started by Professor David Magnusson at the Department of Psychology,
Stockholm University, in mid-1964 (Magnusson et al., 1975; Magnusson, 1988). Professor
Lars R Bergman is the current scientific leader. The broader focus of this study was on how
individual and environmental factors, and also their interactions, influence psychosocial
development over the lifespan. The so-called interactionist perspective on development has
guided research within the IDA program (Magnusson, 1988).
The participants come from the medium-sized Swedish town of Örebro. At the outset of
the investigation in 1965, Örebro had about 80,000 inhabitants. It has expanded over time. In
1985 there were approximately 120,000 persons living in Örebro and in year 2000 130,000.
Given the population of Sweden (currently 9 million), Örebro is a relatively large city by
national standards. The economy of Örebro was at the initiation of the study dominated by a
prominent footwear industry, but has diversified over time into engineering plants, printing
shops, and a food and paper industry. Additionally, Örebro is home to a large hospital and
university. A city-based comparison (Stattin, Magnusson, & Reichel, 1986) showed that
Örebro had rates of crime comparable to similar-sized cities in Sweden (50,000-100,000
inhabitants), and moderately lower rates of crime compared with Sweden’s largest cities (i.e.,
Stockholm, Göteborg, Malmö).
The first data were collected when the participants were 10 years-old and data gathering
is still in progress. The target sample comprised all children who in 1965 attended normal
grade 3 schooling in public compulsory school. There were no private schools in the
community at this time. The total number of children was 1,027. Of those, 510 were girls and
517 boys.
Thus far, data have been collected at ages 10, 13, 15, 16, 17, 18, 27, and 43 for the female
participants. The sample size has grown over time. This is because girls who moved into the
42
43
29
community entered the study while girls who left the community did not leave the study. At
the last data collection, in 1998, 682 women participated. At age 27, participants filled out
postal questionnaires containing questions concerning childbirth, education, work, and
drinking habits. Of those who participated in adolescence, 90% took part in the follow-up at
age 27. When the women were 43 years of age, in 1998, they were interviewed about various
aspects of their adjustment situations. The participation rate was 89%.
Extensive data have been collected from different sources: from the children themselves
information was collected about, for instance, intelligence, school performance, adjustment to
school, anxiety, psychosomatic symptoms, bullying, and vocational preferences; from the
teachers information was collected about, for instance, ratings of aggression, motor
restlessness, lack of concentration, and certain symptoms; from the parents information was
collected about, for instance, education and vocation, conditions of living, family situation in
general and problems with the child; from peers information was collected about, for instance,
social relations; test information was collected about, for instance, achievement and
intelligence; register information was collected from official records about, for instance,
school grades and number of hours absent from school (Magnusson et al., 1975).
3.1.2 Sample II. Sample II was used for Paper I. This sample comes from a Swedish
longitudinal study that was started in the fall of 2001 by researchers at the Center for
Developmental Research (CDR) at Örebro University (the 10 to 18 study). The main focus of
the study was to understand the risk and protective factors and etiological mechanisms behind
antisocial, delinquent behavior. “10 to 18” is a 6-year longitudinal investigation that has been
conducted in one entire community, Köping, in central Sweden. The community has around
26,000 inhabitants. The unemployment rate is the same as the Swedish average and the
average income is slightly lower than the national mean. This particular community was
chosen because it met several research criteria. First, it has its own high school, so youths
remain in the community’s school system over the entire adolescent period. Second, it does
not offer youth easy access to neighboring towns, so most youths have their leisure time peer
associations inside the community. Third, it has a population of a size that enables all
potentially influential peers to be included in the study. The community harbors 13 schools
and 158 classes in the age range covered. Professor Margaret Kerr and Professor Håkan
Stattin are the project leaders.
The participants were all the children and adolescents in the town who were between the
ages of 10 and 18 (N = 3,000), or more specifically all children in grades four through 12.
They were followed annually from 2001 to 2007. Participants were recruited by asking all
youth in grades four to 12 to participate in the study. Parents were informed about the study
beforehand by mail and meetings in the community. They received a pre-stamped postcard
that they could return if they did not want their child to participate. One per cent did so. Thus,
active consent was obtained from the children and passive consent was obtained from the
parents. Data collections were performed during normal school hours. Participants were asked
44
43
30
to name their peers; because virtually all youth aged 10 to 18 took part in the study (90% and
over), there are self-reported, independent data from both the target participants and their
peers. The benefit of this design is that the peer data are not susceptible to reporter or
reporting bias, which often inflates similarity (Iannotti, Bush, & Weinfurt, 1996).
For Paper I, data from girls in grades 8 and 9 during the school year 2001-2002 were
used. During this school year there were 3,174 participants aged 10 to 18 in the community’s
schools in total. Data were collected twice (in fall and spring) separated by 4 months. At the
first data collection, of all 334 girls listed at school in grades 8 and 9, 309 (93%) were present,
and at the second 302 (90%). Information was obtained through self-report questionnaires that
were filled out by youth during school hours.
3.1.3 Sample III. Sample III was used for Paper I. This sample comes from a Slovakian
study, which was conducted in the fall of 1995 by researchers at the Institute of Experimental
Psychology, Slovak Academy of Sciences, in Bratislava. Professor Zdena Ruiselova was
responsible for the data collection. The participating 15-year-old girls took part in broad
surveys of self-experienced symptoms and adjustment problems. The cohort consisted of
freshmen in all high schools in a town (Banská Bystrica) of about 90,000 in central Slovakia.
There were 1,018 girls registered in the town for the spring term. From among these, data
were collected for 955. Hence, the participation rate was 94%. The average age of the girls
was 14 years and 5 months. Trained test leaders were used during data collections, and the
students answered the questions in their regular classrooms. The procedure was identical to
that used with samples I and IV.
3.1.4 Sample IV. Sample IV was used for Paper I. This sample came from a study with
a focus on exploring self-experienced symptoms and adjustment problems among 15-year-old
girls. Professor Håkan Stattin is the project leader. This study was designed to be a repeat of
the IDA study (see Sample II). It was performed as an identical replication of the original
study. The study was conducted in the same community, Örebro. It comprised all the
adolescent girls attending grade 8 in this community in the 1995-96 school year. Hence, it is
referred to as the “95-96 study”. Data collections were made on two occasions – November
1995 and May 1996 – just as in the earlier IDA study. This sample included 602 girls; 529
(88%) of them were present and filled out the questionnaire at the November data collection.
Trained test leaders were used, and the pupils answered the questions in their regular
classrooms.
3.1.5 Sample V. Sample VI was used for Paper I. This sample comes from a short-term
longitudinal study that started in the fall of 1998 and ended in the spring of 2000. The project
is normally referred to as the “Short-term longitudinal study”. Professor Margaret Kerr and
Professor Håkan Stattin were the project leaders. The focus of this study was on testing the
causal links in a model of family communication processes, on discovering how children and
parents react to each other over time, on how those reactions change their relationships, and
on how their relationships are linked to delinquency and other forms of adjustment.
44
45
31
The participants came from the same medium-sized Swedish town, Örebro, which is
described under Sample I. Information was collected from students, parents, and teachers. The
participants were in grade 8 when the study began. The target sample comprised 1,279
students (48% boys; 52% girls). Of the 1,279 students, 1,186 (93%) with a mean age of 14.4,
were present on the day of the first data collection, and answered the questionnaires. The data
were collected during normal school hours. Parents were informed about the study beforehand
by mail. They received a pre-stamped postcard that they could return if they did not want their
child to participate. Twelve parents did so. In the spring of 2000, approximately 18 months
later, 1,057 out of the 1,279 (83%) students participated in data collections that were
performed in class rooms, again by using self-report questionnaires. At this time, the
participants were attending 9th grade. Eighty-seven per cent of the participants were born in
Sweden.
3.1.6 Sample VI. Sample VI was used for Paper IV. This sample came from a Swedish
longitudinal study that was started in the mid-1950s by researchers at the Clinic for the Study
of Children’s Development and Health at the Karolinska Hospital in Stockholm. The study
was part of a larger international investigation that was organized by Centre International de
l’Enfance in Paris. Professor Håkan Stattin is the current project leader. The focus of the study
is on somatic, psychological, and social development. However, it also comprises
information, among other things, on parent-child relationships, sexual development, sleep,
social and economic circumstances, child behavior, school performance, criminality, and
substance use.
Participants were recruited by asking every fourth pregnant mother who registered at the
Solna Prenatal Clinic in (a Stockholm suburb) from April 1955 to April 1958 to participate in
a longitudinal pediatric study. Hence, the study is referred to as the “Solna Study”. Because
virtually all Swedish pregnant mothers receive prenatal care, the selected women were
assumed to be representative of Swedish mothers living in urban areas at this time. Seventeen
mothers moved out of the community during their pregnancy, and seven mothers were unable
to participate due to various reasons. In such circumstances, the next fourth mother was
invited to participate in the study. The number of mothers who agreed to participate was 198.
Six withdrew because of abortion and four because of stillbirths, while five were excluded
because of low childbirth weights (under 2,000 grams). Further, a pilot group of 29 children,
who were contacted either before or just after birth, were added to the study. In total, 183
children plus the 29 children from the pilot group made up the sample. Of these, 122 were
boys and 90 girls. Data from the 90 girls were used for Paper IV.
The children were assessed six times during their first year of life, twice during their
second year, and once per year up to age 18 thereafter. Data collections in adulthood took
place at ages 21, 25, and 37. Up to 18 months, participants were examined as close as possible
to their birthdays to avoid age effects. From 18 months and onwards, this was within four
weeks before or after their birthdays. Eighty-five and 90 per cent participated at ages 25 and
46
45
32
37, respectively. Information was collected through somatic registrations, medical
examinations, interviews, inventories, ratings, objective tests, sociometric methods, and
projective techniques. Multiple investigations have shown the sample to be representative of
people living in Swedish urban communities (e.g., Karlberg, Engström, Lichtenstein, &
Svennberg, 1968).
46
47
48
47
Used
I
I
I
IV
III
IV
V
VI
study
The Solna
study
longitudinal
term
The Short-
study
90
618
Longitudinal
Longitudinal
Cross-sectional
1955
1998
1995
1995
2002
1965
1995
2000
1996
1996
On-going
On-going
Ended
Age at
0-37 years
14-16 years
14-15 years
14-15 years
14-16 years
10-43 years
collections
data
Banská
Sweden
Köping,
Sweden
Örebro,
Place
Sweden
Stockholm,
Sweden
Örebro,
Sweden
Örebro,
Slovakia
The “95-96”
529
Began
(year)
study
Cross-sectional
Longitudinal
Longitudinal
Design
Bystrica,
955
334
522
Ns
Slovakian
The
study
The 10 to 18
project
III
I
The IDA
Name
I, II, &
Paper
in
II
I
Sample
Table 1.
Characteristics of the six samples used in this dissertation.
questionnaires
reports, and self-report
Physical examinations, parent
Self-report questionnaires
Self-report questionnaires
Self-report questionnaires
and peer questionnaires.
Parent reports, self-reports
questionnaires
Parent reports and self-report
Procedure
Growth development
adjustment
Family processes and
and adjustment problems
Self-experienced symptoms
and adjustment problems
Self-experienced symptoms
Antisocial development
adaptation in context
Individual development and
Main focus
34
3.2 Measures
3.2.1 Paper I. Study 1. Normbreaking behavior, used for the Swedish 1971-96
comparison, was a scale aggregated from eleven items in the so-called Symptoms
Questionnaire about shoplifting, alcohol abuse, signature forgery, pilfering, using hashish,
vandalism, evasion of payment, physically hitting somebody, harassment, and loitering in
town in evenings (Magnusson et al., 1975). Responses were on a 5-point scale: (1) no, (2)
once, (3) 2-3 times, (4) 4-10 times, and (5) more than 10 times. Alpha reliability for the scale
was .75 in 1971 and .84 in 1996.
Normbreaking behavior, used for a cross-cultural comparison, was taken from the Norm
Inventory (Magnusson et al., 1975). The instrument covered eight norm-violating activities: at
home (ignoring parents’ prohibitions, staying out late without permission); at school (cheating
on an exam, playing truant); and, during leisure time (smoking hashish, getting drunk,
shoplifting, loitering in town in the evenings). For each of these, participants reported on their
own evaluations. The answers were given on 7-point scales with the alternatives: (1) very
foolish, (2) foolish, (3) rather foolish, (4) not really OK, (5) rather OK, (6) OK, and (7) quite
OK. The measure of the girls’ own norm violation consisted of answers to the question how
many times the girls had performed each of the eight norm-violating behaviors. Answers were
given on 5-point Likert scales with the alternatives: never (1), once (2), 2-3 times (3), 4-10
times (4), and more than 10 times (5).
Behavioral intentions in connection with normbreaking. For each of the eight normviolating activities in the Norm Inventory, descriptions of concrete instances were
constructed. Situations depicting the different types of norm-violating activities were
presented, and the girls were asked about their own behavioral intentions to breach respective
norms (“What do you think you would do in this situation?”) on 7-point scales, such as:
would definitely not cheat (1), would probably not (2), would perhaps not (3), uncertain what
to do (4), would perhaps (5), would probably (6), and would definitely (7). The eight normviolating activities for each of the different norm dimensions (evaluations, behavioral
intentions, and own norm-violating behaviors) were aggregated into broader scales. Alpha
reliability for the normbreaking behavior (own behavior) scale was .84 for Swedish girls and
.81 for the Slovakian girls. For the two other scales (evaluations, behavioral intentions), the
alphas varied between .84 and .90 for the Swedish and Slovakian samples. Hence, very small
departures from homogeneity for the scales were found between girls in the two countries.
Romantic relationships with boys. Three items measured how advanced the girls’
relationships with boys were. One was about perceived sexual maturation: “Do you feel
sexually more experienced than your age-mates?” The response options were: never (1),
seldom (2), sometimes (3), rather often (4), and very often (5). Another question was about
intercourse: “Have you had sexual intercourse?” The response options were: no (1), yes, once
(2), and yes, several times (3). The third question was about having a steady boyfriend: “Have
you now or have you had a steady relationship were a boy?” The response options were: have
48
49
35
never had, don’t want to have (1), have never had, but want to now (2), have had earlier, but
not now (3), have now, but not earlier (4), and have now, and have had earlier (5). The final
item was transformed into a trichotomized measure in the calculations. Scale values 1 and 2
were coded 1, value 3 was coded 2, and values 4 and 5 were coded 3. Alpha reliability for this
3-item scale on romantic relationships with boys was .70 in 1971 and .74 for the 1996 sample.
Pubertal timing. Information on age at menarche was requested in the Symptoms
Questionnaire. It was originally coded in 6 categories: before age 10 (1), between 10 and 11
(2), between 11 and 12 (3), between 12 and 13 (4), after age 13 (5), and have not yet had my
first period (6). In all studies in Paper I, a condensed four-scale measure was used,
differentiating between girls who had their menarche: before 11 years (1), between 11 and 12
(2), between 12 and 13 (3) and 13 and later (4). The coding highlights variations in early
timing – commensurate with the theoretical interest – rather than covering the whole age
variation of pubertal timing. Using the IDA sample, Stattin and Magnusson (1990) compared
this self-report measure with bone ossification and weight and height measures at the age of
13 years. They found strong support for the validity of self-reported age at menarche as a
measure of girls’ pubertal timing.
Study 2. School and free-time peer groups. Girls were asked whether they belonged to a
group (containing at least three persons) that hanged out together at school. If they mentioned
such a group, they were asked to list the names of all the youngsters who belonged to this
group. They could mention up to ten school peers in total. The participants were also asked if
they belonged to a group that hanged out together in their free time. If they belonged to such a
group, they were asked to list the names of the persons who belonged to this group. They
could mention up to ten free-time peers in total.
Age and delinquency of peers. The peers in the school group and the peers in the freetime group were identified through class lists. For each peer mentioned, we added their
gender and age (according to year and month). Because these peers in most cases answered
our questionnaires, we computed a summary score of delinquency based on their self-reported
delinquent activities. Thus, delinquency (and also age and gender) in peers was not
ascertained through reports of the participants who mentioned these peers, but through selfreports of the peers themselves. The delinquency measure was composed of 22 questions.
Youths answered these questions about their behavior over the past year: “Have you taken
things from a store, stand, or shop without paying?”, “Have you been caught by the police for
something you have done?”, “Have you, on purpose, destroyed things, such as windows,
street lights, telephone booths, benches, gardens, etc.?”, “Have you taken money from home
that was not yours?”, “Have you been part of painting graffiti, or writing with markers or
spray paint on, For instance, a sidewalk?”, “Have you participated in breaking into a home,
shop, stand, storage building or other buildings with the intention of taking things?”, “Have
you stolen something from someone’s pocket or bag?”, “Have you bought or sold something
that you knew or thought had been stolen?”, “Have you taken a bicycle without permission?”,
50
49
36
“Have you threatened or forced someone to give you money, cigarettes, or anything else?”,
“Have you taken part in a fight on the street in town?”, “Have you carried a weapon (e.g.,
brass knuckles, club, knife, switchblade, or some other weapon)?”, “Have you participated in
taking a car without permission?”, “Have you taken a moped, motorcycle, or Vespa without
permission?”, “Have you been part of hitting someone so that you believed or knew that he or
she needed to be treated at the hospital?”, “Have you intentionally hurt someone with a knife,
switchblade, brass knuckles, or some other weapon?”, “Have you taken part in threatening or
forcing someone to do something that he or she didn’t want to do?”, Have you snuck out
without paying (e.g., at movies, a café, on the train or bus, or somewhere else?”, “Have you
taken part in stealing something from a car?”, “Have you drunk so much beer, liquor, or wine
that you got drunk?”, “Have you smoked hashish (marijuana, cannabis)?”, and “Have you
used any drugs other than hashish (marijuana, cannabis)?”. Alpha reliability for a scale
summing these items was .92.
Sexuality and steady relationships. The participants were asked “Have you had sexual
intercourse?” and answered on a three point scale ranging from (1) no, never (2) yes, once, to
(3) yes, several times. They were also asked if they had a steady boyfriend relationship,
answered on a 5-point Likert scale: (1) have never had, and do not want to, (2) have never
had, but can think about it now, (3) have had, but not now, (4) have now, but have not had
earlier, and (5) have now and have had earlier. The measure was recoded as reported in
Study 1
Activities the last month with free-time peer groups. Participants were asked what
activities they had undertaken with their free-time peers the last month. They answered
separately (1. no, 2 yes, once, and 3. yes, several times) for the following 11 activities:
Studied, Watched TV/videos, Played hookey, Talked on the telephone/chatted, Shoplifted,
Used the computer, internet, or played TV videogames, Drank alcohol until you were drunk,
Kept secrets from parents, Talked about things that are illegal, Went out looking for boys, and
Did other things, apart from shoplifting, for which one could get caught by the police.
Study 3. Youth center participation. The 1998 cohort of participants responded to a
question of whether they usually attended youth centers in the town (1. No, 2. Yes); if they did
so, they were asked how many evenings per week they typically were there. The response
options for the 1970 cohort were on a 5-point scale, ranging from no participation to
participation every evening of the week. Note that an adolescent has to be 13 years of age to
gain entrance to these centers.
The 1970 cohort
Romantic relationships with boys. Three items measured how advanced the girls’
relationships to boys were; the items were identical to those described in Study 1.
50
51
37
Problem behavior was the same scale in the Norm Inventory previously described under
Study 1.
Peer involvement at age 13 was a composite measure based on three questions “How
many peers do you have that you are together with really often after school?” “How many
evenings per week do you usually meet your peers?” and “How many evenings per week do
you usually spend at home?” (reversed). The inter-item correlation for these three measures
was acceptable, at .34, but the alpha for the three-question peer involvement scale was low,
at.64.
The 1998 Cohort
Romantic relationships with boys. The girls answered 3 questions on romantic
relationships with boys: “Do you have, or have you ever had a steady boyfriend?”, Have you
had sexual intercourse?”, and “If you have a boyfriend/girlfriend, how often do you see each
other?“. Alpha reliability was .75 (see Study 1).
Delinquency. The 1998 cohort responded to a delinquency questionnaire. They answered
15 questions about whether they had engaged in certain behaviors during the past year.
Responses were on a 5-point scale ranging from “never” (1) to “more than 10 times” (5). The
questions were about: shoplifting; being caught by the police for something they had done;
vandalizing public or private property; taking money from home; creating graffiti; breaking
into a building; stealing from someone’s pocket or bag; buying or selling stolen goods;
stealing a bike; being in a physical fight in public; carrying a weapon; stealing a car; stealing a
moped or motorcycle; using marijuana or hashish; and using other drugs. Alpha reliability
was .72.
Peer involvement. The participants reported how many nights per week they usually
socialized with friends. The response options ranged from one to seven days per week.
The 1970 Cohort – follow-up
Romantic relationships with boys. At the average age of 14 years and 10 months (grade
8) girls in the sample reported whether they had sexual intercourse on a 3-point scale
(alternative responses: 1. no, 2. yes, once, and 3. yes, several times).
Youth recreation-center attendance. At age 13 (grade 6), the girls reported the frequency
with which they attended any of the youth recreation centers operating in the town of Örebro.
Center attendance was coded on a 5-point scale ranging from no participation, to participation
every evening of the week (Mahoney, Stattin, & Magnusson, 2001).
Registered delinquency. Information about criminal offense was obtained from official
registers. For underage offending (below age 15), information about juvenile offending was
obtained from local welfare authorities. Information about registered offenses between age 15
and age 30 were obtained from police registers. Only three of the girls were underage
52
51
38
offenders. At age 30, 38 (7.5%) of the 510 girls with data on age at menarche were registered
by the police for one or several criminal offenses.
Study 4. Social Behavior. Information about behavioral problems came from teacher ratings
and peer nominations when the participants were 10 and 13 years of age. Data included
teacher ratings concerning aggressiveness, concentration problems, motor restlessness (the
latter two, measures that have been used as global indicators of hyperactive behavior),
disharmony, and school fatigue. Teacher ratings were made on a 7-point scale, where higher
values indicate more of the rated construct (Magnusson et al., 1975). It should be noted that in
most cases the teachers who made these ratings had had the opportunity to observe the girls
over three years. The teachers who had been the head teachers from age 7 to 10 made the age
10 ratings, and the teachers who had been the head teacher from age 10 to 13 made the age 13
ratings. In almost all cases, the teachers who made the ratings at age 10 were different from
the teachers who did the ratings three years later.
Intelligence. An intelligence test, the Westrin Intelligence Test (WIT III), was
administered to the participants at age 14. The test contains four subtests measuring inductive
ability, verbal comprehension, deduction, and spatial ability. The manual reports a split-half
reliability of .93.
Grades. Grades in Swedish and Mathematics were available for the girls at the ages of 10
and 13. The head teachers provided them on 5-point scales, with 5 being the best grade. We
aggregated the grades over the two subjects.
Popularity in the class. To assess popularity at ages 10 and 13, students were asked to
imagine that they were being transferred to another classroom, but that not all of their
classmates could go to the new room (c.f., Magnusson et al., 1975). They were then asked to
rank all of their classmates in the order in which they would like them to move along. The
students were asked to rank girls and boys separately. The rankings were then combined
across all informants and standardized to form a continuum of popularity. A constant of 3 was
added to avoid negative scores. Higher scores on this measure reflect a greater number of peer
nominations (labeled popularity).
Perceived popularity. The participants were also asked about how they, themselves,
would be scored according to nominations of their peers in the class. This measure, perceived
popularity, was a self-report measure on where in their peer ranks the girl thought that she
was placed.
Poor peer relationships. Four items made up a scale measuring poor peer relationships at
school at age 10; “Do you enjoy the peers in your class?”, “Do you have fun with peers during
the breaks?”, “Can you play with those you want to play with during the breaks?”, and “How
many of your classmates do you like?” Alpha reliability for the 4-item scale was .65 (interitem correlation was .31). Six items made up a similar measure of poor peer relationships at
school at age 13: “Do you have a really good friend in your class?”, “Do you enjoy your
52
53
39
classmates?”, “Do you have a good time during the breaks?”, “Do peers pick on you during
the breaks?”, “How many of your classmates do you like?”, and “Do you get along well with
your classmates?”. Alpha reliability for the age 13 measure was .78 (inter-item correlation
was .37).
Parental concerns. When the girls were 10 years of age, their parents were asked if they
had any concerns in the following domains with respect to the child: pocket money, TVwatching, doing homework, curfew time, peer relationships, smoking, and alcohol or sniffing,
and disobedience. The response scale was a dichotomized measure: Yes or No. Alpha
reliability for the scale combining these eight domains was .66. A similar parental concern
scales was made up at age 13, with an alpha reliability of .74. It involved seven questions to
parents about the child’s pocket-money, TV-watching, doing homework, curfew time, peer
relationships, disobedience, and going to bed.
Thrill seeking was a subscale of the Youth Psychopathic traits Inventory (YPI), an
adolescent self-report measure of psychopathic personality traits (Andershed, Kerr, Stattin, &
Levander, 2002). The following 9 items, reported on a 4-point scale ranging from (1) does not
apply , to (4) apply precisely, are included: ”I get bored quickly by doing the same thing
over,” ”I get bored quickly and I like change,” ”I like to do things just for the thrill of it,” ”I
like to do exciting and dangerous things, even if it is forbidden or illegal,” ”I like people that
are involved in exciting and unexpected activities,” ”I have an unusually great need for
change,” ”I’m drawn to places where exciting things happen,” ”I like having ’lively and
active’ surroundings,” and ”To be on the move, traveling, change, and excitement – that’s the
kind of life I like.” The alpha reliability was .82.
Impulsivity was a subscale of the Karolinska Scales of Personality (KSP). The following
six items were answered on a 4-point Likert scale (1 = never agree, 2 = rarely agree, 3 =
sometimes agree, and 4 = always agree) ”I have a tendency to act on the spur of the moment
without really thinking ahead,” ”I usually get so excited over new ideas and suggestions that I
forget to check if there are any disadvantages,” ”I often throw myself too hastily into things,”
”I am a very conscientious person (rev.),” ”I usually talk before I think,” and ”I consider
myself an impulsive person.” Alpha reliability was .74.
The risk taking scale contained items from different self-report instruments to measure
psychopathic traits. Unless otherwise stated, youths responded on a 5-point scale from
strongly disagree (1) to strongly agree (5). The items were as follows: ”I would do almost
anything on dare;” “I dare to do things that others do not” (response scale: 1 = true, 2 = false);
”I enjoy traveling in cars at high speed;” ”I have often done something dangerous just for the
thrill of it;” ”I enjoy taking chances;” ”I would never do something dangerous just for the
thrill of it” (reversed); ”I enjoy gambling for large stakes;” ”I enjoy doing wild things;” and ”I
would do almost anything on a dare” versus “I tend to be a fairly cautious person”
(dichotomized measure). The alpha reliability was .83.
54
53
40
Normbreaking. The measure of normbreaking behavior was taken from the Norm
Inventory at the average age of 14 years and 10 months (Magnusson et al., 1975), which
consisted of eight norm-violating activities: at home (ignoring parents’ prohibitions, staying
out late without permission), at school (cheating on an exam, playing truant), and during
leisure time (smoking hashish, getting drunk, shoplifting, loitering in town in the evenings).
See above.
Romantic relationships with boys. Four items measured how advanced the girls’
relationships to boys were at age 14 years and 10 months. Girls were asked about their
perceived sexual maturation, perceived popularity among boys, sexual intercourse, and steady
boyfriend relationships; the alpha for this scale was rather low, at .61, with an inter-item
correlation of .29.
3.2.2 Paper II. For this paper, age at menarche was used as the measure of girls’
pubertal timing. Girls were asked about their ages at menarche in grade 8 (the age of 15). Five
girls (1%) had their first menstruation before the age of 10, 46 (9%) between 10 and 11 years,
117 (22.9%) between 11 and 12 years, 207 (40.6%) between 12 and 13, and 106 (20.8%) at
the age of 13 or later; 29 girls (5.7%) had not had their first menstruation at the time of data
collection. Thus, the typical age of menarche among the girls in the sample was between 12
and 13 years. In addition, we created a categorical measure. Girls, who had their menarche
before age 12, were assigned to an early developers group. Girls, who had their menarche
after age 12, were assigned to a later developers group.
Second, I was interested in girls’ self-perceptions of their maturity. In grade 8, in a
questionnaire that measured different aspects of peer and parent relations, girls were asked if
they felt more or less mature than their classmates, on a 5-point scale with the scale values:
(1) not at all as mature as most others; (2) not really as mature as most others; (3) about as
mature as most others; (4) somewhat more mature; and (5) much more mature than most
others. In addition to this continuous measure, we created a dichotomized measure dividing
the girls into groups of girls who felt more mature than their peers (scores 4 and 5) and girls
who did not feel more mature than their peers (scores 1, 2, and 3). This is a measure of girls’
subjective perception of their maturity. When asked to describe the concept of maturity,
adolescents rarely talk about physical aspects. Instead, the majority focuses on psychosocial
or “genuine” aspects of maturity (Tilton-Weaver et al., 2001). Thus, it is more likely that girls
were thinking of their behaviors rather than their physical maturity when asked this question.
Third, I was interested in girls’ level of engagement in heterosexual relations (referred to
in Paper II as “romantic relationships with boys”). Four items measured how advanced the
girls’ heterosexual relationships with boys were in grade 8. Girls were asked about their
perceived sexual maturation, “Do you feel sexually more experienced than your age-mates
(1=never, 2=seldom, 3=sometimes, 4=rather often, and 5=very often), intercourse, “Have you
had sexual intercourse?” (1=no, 2=yes, once, and 3=yes, several times), attitude towards
intercourse (1=very foolish to 7=totally OK), and steady boyfriend relations “Do you have or
54
55
41
have you had a steady relation to a boy?” (1=have never had, don’t want to have, 2=have
never had, but I want one now, 3=have had earlier, but not now, 4=have now, but not earlier,
and 5=have now, and have had earlier). Since the items had different numbers of response
alternatives, we standardized them before computing the composite scale. Alpha reliability for
this scale was .73. The girls were then further divided into a group with little experience of
heterosexual relationships and a group with a lot of experience of heterosexual relationships,
using a median split.
I divided the girls into eight subgroups based on the dichotomous measures of pubertal
timing, perceived maturity, and heterosexual relations. The first group consisted of laterdeveloping girls who did not feel more mature than peers, and who had little experience of
heterosexual relationships (the Reference Group; n = 131). The second group contained laterdeveloping girls who did not feel more mature than their peers, and who had a lot of
experience of heterosexual relationships (the Late Relationship-Only Group; n = 87). The
third group consisted of later-developing girls who felt more mature than their peers, and who
had little experience of heterosexual relationships (the Late Mature-Only Group; n = 22). The
fourth group contained later-developing girls who felt more mature than their peers, and who
had a lot of experience of heterosexual relationships (the Late Precocious Group; n = 30). The
fifth group consisted of early-developing girls who did not feel more mature than their peers,
and who had little experience of heterosexual relationships (the Premature Group; n = 32).
The sixth group consisted of early-developing girls who did not feel more mature than their
peers, and who had a lot of experience of heterosexual relationships (the Early RelationshipOnly Group; n = 50). The seventh group consisted of early-developing girls who felt more
mature than their peers, and who had who had little experience of heterosexual relationships
(the Early Mature-Only Group; n = 13). The eighth group consisted of early-developing girls
who felt more mature than their peers, and who had a lot experience of heterosexual
relationships (the Early Precocious Group; n = 41).
I employed a scale comprising seven items to measure problematic parent relations.
Examples of items were: “Do your parents listen to you?”, and “Are your parents
disappointed with you?” The scale was internally consistent ( = .87). Girls’ adjustment to
school was examined in grade 8. This was measured on a scale comprising nine items. Some
of the items were: “Are you satisfied with your school work?”, “Is it important to perform
well in school?”, and “Are your teachers fair to you?” Alpha reliability for this scale was .83.
How many peers of different sorts the girls had and how often they associated with peers was
explored in grade 8. Girls reported the number of peers they had and how many days per
week they met with their peers (peer regularity). Girls’ normbreaking was a scale aggregated
from eleven questions about shoplifting, signature forgery, pilfering, using hashish,
vandalism, evasion of payment, physical aggression, harassment, and loitering, which were
posed in grade 8 (Magnusson et al., 1975). The 5-point scale had the following response
options: 1 = no, 2 = once, 3 = 2-3 times, 4 = 4-10 times, and 5 = more than 10 times. Alpha
56
55
42
reliability for the scale was .73. I wanted to explore girls’ alcohol drinking separately from
other types of normbreaking behaviors, because of the heightened awareness of adolescent
girls’ drinking behavior. Girls were asked whether they had ever been drunk. The response
scale was the same as for the items tapping normbreaking.
At age 26, participants filled out postal questionnaires covering educational-vocational
career, family life, working conditions, social networks, and free-time activities. Among other
things, the women were asked about their family lives, educational attainments, work-life, and
alcohol-drinking. First, it was asked if they were married/cohabitant or not. The response
options were (1) no or (2) yes. Participants were also asked how many children they had.
Response options were: none, one, two, three, four, five, six, or seven. The mean number of
children was .66. Participants also reported on the age of their oldest child. Females reported
on their level of education on an 8-point scale, ranging from no education past compulsory
school (ending at age 16) to at least four years of university studies (at age 26). In addition,
they were asked whether they were in employment, and they answered an open question about
their types of employment at the age of 26. From this information, job descriptions were coded
on an 8-point work position scale according to the minimum education required for each kind
of employment. Furthermore, women were asked about their alcohol drinking habits. They
reported on how much alcohol they drank at the most during one occasion and how often they
drank alcohol. Information about criminal offenses was obtained from official registers. For
juvenile offending (below the age of 15), information was obtained from local welfare
authorities. Information about registered offense between the age of 15 and the age of 30 were
obtained from police registers.
At age 43, the women reported on the number of children they had given birth to. This
measure includes children still living with the participant and children who had already left
their parental homes. The women reported on whether they lived with a husband or had a
cohabitant. Participants’ educational levels were investigated. Answers were coded on a 7point scale ranging from no education past compulsory school, to at least a university degree.
There was also an “other” response option. This option was omitted from the analyses. The
women reported on their partners’ highest educational level on 10-point scales, where 10
indicated the highest education level (a university degree).
3.2.3 Paper III. Height. Females’ heights were measured at ages 13 and 43 during
physical examinations.
Weight. Females’ weights were measured at ages 13 and 43 during physical
examinations.
Body Mass index (BMI). The formula for calculating BMI is weight (in kilograms)
divided by height (in meters) squared. The higher BMI a woman has, the stouter she is. There
are strong correlations between BMI, fat mass, and percentage of body fat in children,
adolescents, and adults (Daniels, Khoury, & Morrison, 1997; Pietrobelli, Faith, Allison,
Gallagher, Chiumello, & Heymsfield, 1998).
56
57
43
Waist circumference in centimeters was measured during physical examinations at age
43.
Hip circumference in centimeters was measured during physical examinations at age 43.
Physical fitness. At age 43, the women graded their physical fitness as their ability to
walk, jog, or run 2 km. They were asked if they agreed to one of the following statements: (1)
I cannot walk 2 kilometers without a rest, (2) I can walk 2 kilometers without a rest, (3) I can
jog 2 kilometers if I stop and rest a couple of times, (4) I can jog 2 kilometers without a rest,
(5) I can run 2 kilometers at a good speed, if I can stop and rest a couple of times, (6) I can
run 2 kilometers at a good speed without a rest, and (7) I can run 2 kilometers at high speed
without rest.
Dieting. When the women were 43 years-old, they were asked “How often do you diet?”
Answers ranged from (1) No, never to (3) Yes, often or almost always.
Quality of life. Positive affect was measured using a scale comprising nine questions
concerning women’s feelings of being happy, satisfied, forward looking, relaxed, calm and
peaceful, etc. during the past month (alpha = .92). Positive affect was measured using the
scale “General positive affect” in the Mental Health Inventory (Veit & Ware, 1983). Selfimage was a scale that measured positive self-perception. It contained four questions, like “I
am proud of the type of person I have become” (alpha = .72). Family satisfaction was tapped
by one item asking subjects “Are you happy with your family life?” Work satisfaction was
measured on a scale comprising four items, such as “I am satisfied with my current work
position” (alpha = 0.84). Women reported on their leisure satisfaction by answering a single
question about how happy they were with their leisure time. Life satisfaction, finally, was
assessed using two questions, “Are you satisfied with your life” and “How do you like your
current life?” The correlation between these two items was 0.35.
3.2.4 Paper IV. Age at menarche. From age 10 to 16, girls were asked whether their
menstrual bleeding had occurred. This was done every third month during telephone
interviews.
Body mass index (BMI). Trained staff collected information on height (in centimeters)
and weight (in kilograms with an accuracy of .1 kilograms) during medical examinations
when participants were 8, 25, and 37 years-old. We used this information to calculate
participants’ BMI. Females’ BMI scores were computed using the same formula as in Paper
III.
Appetite. At each year between ages 6 to 16, parents were asked if their child had had a
good appetite. Answers were given on 5-point scales. Response options were: (0) never, (1)
seldom, (2) sometimes, (3) often, and (4) always. A higher score indicated a higher appetite.
We scored the girls’ appetites one year before their menarche in an attempt to compare girls’
appetites when they were at approximately the same physical maturity level, and thus avoid
confounding pubertal maturation and appetite. The average year-to-year correlation for the
appetite measure was r = .59, with rs ranging from .45 to .73.
58
57
44
3.3 Statistical analyses
The majority of the calculations for the studies were performed using SPSS 13.0 or 14.0. In
Paper II, I also used the CFA module in the SLEIPNER 2.0 software (Bergman & El-Khouri,
1998). This was to enable a configural frequency analysis (CFA) to test developmental
trajectories (Lienert & zur Oeveste, 1985). The CFA module uses the exact binomial test to
perform configural frequency analyses. Frequencies with which configurations of certain
categorical variables are observed are examined. A CFA is used to test special cells, in crosstabulations or contingency tables, in which expected and observed frequencies for each
configuration are compared. An outcome is labeled a “type” when the observed frequency in
a cell significantly exceeds the expected frequency. Conversely, when the expected frequency
in a cell significantly exceeds the observed frequency, the outcome is labeled an “antitype”.
Thus, this is a person-oriented rather than a variable-oriented approach to analyzing data; it
focuses on characteristics of individuals or subgroups rather than relations between variables
in the total population. A CFA is a recommended analytical tool when adopting a
biopsychosocial approach, where biological, psychological, and social factors are viewed as
deeply intertwined (Lemay, 1999).
Moderation and mediation were tested according to Baron and Kenny (1986). These two
concepts are often confused, and therefore it is important to define them. Moderation concerns
an interaction effect, where the association between two variables (A and B) depends on the
level of a third variable (C). Ge et al.’s (2007) contextual amplification hypothesis builds on
the idea of moderation, in which the link between pubertal timing and adjustment depends on
how adverse or resourceful the context is. Mediation concerns mechanisms, where the
association between two variables (A and C) is explained by a third variable (B), so that A
leads to B and B leads to C.
The level of significance was set at p 0.05.
58
59
45
4. Results
In the section below, I present the specific research questions, hypotheses, and results in each
of the four papers.
4.1 Paper I
The aim of Paper I was to show whether the peer socialization hypothesis satisfactorily
explains the link between female pubertal timing and problem behavior. The peer
socialization hypothesis states that there is a link between early pubertal timing and girls’
problem behavior because early-developing girls start affiliating with older peers and
boyfriends, and through them are channeled into socially advanced behaviors. It is assumed
that the link between pubertal timing and problem behavior should be strongest in contexts
where heterosexual relationships between adolescents are sanctioned and where there is an
over-representation of deviant youth. Paper I reports on four different studies.
4.1.1 The research question of Study 1. The purpose of this study was to compare
two countries – Sweden and Slovakia – which differ considerably with respect to people’s
views on adolescent sexuality, with Slovakia being more conservative and Sweden being
more liberal. Because of this difference, we predicted that if romantic relationships with boys
increase the risk of problem behavior for early developing girls, then the link between
pubertal timing and problem behavior, should be stronger for Swedish than for Slovakian
girls. We predicted that the findings obtained for the Swedish girls would hold for a current
and a 30-year-old sample.
4.1.2 The findings of Study 1. Early pubertal timing was associated with problem
behavior and romantic relationships with boys, and normbreaking behavior was significantly
associated with the three measures of romantic relationships with boys in both Swedish
cohorts. When romantic relationships with boys were controlled for, pubertal timing was not
significantly associated with problem behavior in the 1970 cohort, but associated
significantly, albeit at a low level, in the 1995 cohort. Thus, the mediation model was largely
supported in both cohorts.
No differences were found between the countries with respect to age at menarche. Earlydeveloping girls in Sweden were considerably more experienced with boys than earlydeveloping girls in Slovakia. Pubertal timing was significantly associated with normbreaking
in the Swedish, but not the Slovakian, cohort. Romantic relationships with boys were
correlated with normbreaking among girls in both cohorts.
4.1.3 The conclusion of Study 1. The findings indicate that the conditions favorable
to peer socialization have apparently been met over at least two generations in Sweden. But
the same mechanisms do not seem to operate in Slovakia. Specifically, when young
adolescent girls’ contacts with boys are restricted, early pubertal maturation is not linked to
problem behavior.
59
61
46
4.1.4 The research question of Study 2. The purpose of the second study was to
explore whether chronological age and delinquency of peers differ between early- and latedeveloping girls. Our hypothesis was that we would not find much difference between the
early- and the late-developing girls if the focus was on the peers met only at school, or peers
met both at school and during leisure time. This was because girls are not free to choose the
peers with which they associate at school. For instance, grade-9 girls are the oldest at their
school, and therefore it is impossible for them to associate with older peers at school. We
should, on the other hand, expect that the peers met exclusively in free time to be older and
more delinquent. We used information from all youths in the community; with this wholecommunity design, we obtained information about the social behavior of the girls’ peers,
uncontaminated by the girls’ own perceptions and biases.
4.1.5 The findings of Study 2. Early-developing girls reported having more often
kept secrets from their parents and drunken alcohol to intoxication in company of their freetime peers the last month than the late-developing girls. They also tended to have played
hookey more often with their free-time peers (p < .10). The peers in the free-time group for
the early-developing girls were chronologically older than they were for the late-developing
girls, and the peers in the free-time group were also more delinquent than were the free-time
peers of the late-developing girls. There were scarcely any differences between the early- and
the late-developing girls with regard to the age and delinquency characterization of the group
of peers exclusively met at school and the peers who were met both at school and during
leisure time.
4.1.6 The conclusion of Study 2. Our hypothesis was supported. The peers met at
school did not differ in terms of age or delinquency between girls of different pubertal timing,
but peers met exclusively during free time did differ in this respect, with early-developing
girls having more deviant and older peers who they had met outside school.
4.1.7 The research question of Study 3. If the peer socialization hypothesis is
correct, early pubertal timing should be linked to problem behaviors in leisure settings that
provide affiliation with more advanced or more problematic peer groups, and less in settings
that do not provide access to more advanced peers and behaviors. In the third study, we
explored one example of such a setting: neighborhood youth recreation centers. These centers
have the characteristics that theoretically should enhance the peer socialization effect. Thus, if
the peer socialization hypothesis is correct, then early pubertal maturation should be more
strongly linked to normbreaking among girls who attend these centers than among those who
do not.
4.1.8 The findings of Study 3. The findings for the 1970 and the 1998 cohorts were
similar. Pubertal timing, center attendance, and romantic relationships with boys were all
unique predictors of delinquent behavior in multiple regression analyses. In the group of latedeveloping girls, delinquency was low among those who had little experience of romantic
relationships with boys as well as among those who had much of such relationships. If early-
60
62
47
developing girls had little experience of romantic relationships with boys, their delinquency
was low, but delinquent behavior on their part was much more common if they were heavily
involved with boys.
In agreement with the hypothesis that early maturation should have particularly high
levels of delinquency if they associate with romantic partners and attend youth centers, we
found that among girls who seldom visited youth recreation centers, delinquent behavior was
slightly higher among those who had a lot of experience of romantic relationships with boys
than among those who had little. This was somewhat more pronounced among the earlydeveloping girls. Of those who frequently visited youth recreation centers, delinquency was
not associated with engagement in romantic relationships with boys among the latedeveloping girls. By contrast, for the early-developing group of girls who were frequent
attendees of youth recreation centers, those who had little experience of romantic
relationships with boys showed little delinquent behavior, while delinquent behavior was
considerably more common among the early-developing girls who had a lot of experience of
romantic relationships with boys.
The results were similar when romantic relationships were replaced by peer involvement.
Among early maturers, low peer involvement was associated with very low delinquency
scores, whereas high peer interaction was associated with very high delinquency scores. The
predicted three-way interaction between peer involvement, center attendance, and pubertal
timing appeared in one of the two cohorts. In the 1998 cohort, the three-way interaction was
significant. Peer involvement was related to delinquency for girls who matured early, but not
for those who matured later, and frequent youth center attendance seemed to amplify this
effect. In summary, the main results of these analyses are that particularly high delinquency
scores are found among the early-developing girls who visit the centers often and engage
often with their peers. Thus, the findings are consistent with predictions from the peer
socialization hypothesis and from the idea that the youth centers are a context that can amplify
the peer socialization effect.
Finally, we were interested in the long-term implications of the constellation of pubertal
timing, early romantic relationships, and attending youth centers. We compared the
percentage of police reports among the early-developing girls who did not attend youth
recreation centers frequently in early adolescence and had not had sexual intercourse at an
early age with that among the early-developing girls who either visited these centers often or
had had sexual experiences (see Figure 4.8 in Paper I). Three times as many of the earlydeveloping girls who either had early romantic relationships or who visited youth centers
were registered for an offense by age 35 than was case for the group those who had neither.
None of the late-developing groups showed a prevalence higher than 12.5%. Note that the
percentage of the early-developing girls who showed this particular combination of frequently
attending youth recreation centers and having early sexual experiences was considerably
higher (31.3%).
61
63
48
4.1.9 The conclusion of Study 3. Being an (a) early-developing girl, with (b) a lot of
involvement either in romantic relationships or with peers and (c) visiting youth recreation
centers where there is an over-representation of deviant youth in early to mid-adolescence,
implies a risk factor for adolescent normbreaking and future criminality.
4.1.10 The research question of Study 4. In the fourth study, we examined
whether potential differences in normbreaking behavior in mid-adolescence between early
and later developed girls can be traced back to personality characteristics and individual
differences in problem behavior even at the age of 10, when most girls are pre-pubertal.
4.1.11 The finding of Study 4. There was a small but significant correlation,
suggesting that the earlier girls developed, the more motor restlessness they displayed at age
10. At age 13, teachers rated the early-developing girls as being more restless and as having
more concentration problems. The question was, however, if the associations between
pubertal timing and romantic relationships with boys, and between pubertal timing and
normbreaking, can be explained by early motor restlessness. Age 10 (and age 13) motor
restless was a significant predictor of both age 15 normbreaking and romantic relationships
with boys at age 15. Pubertal timing, though, was a stronger predictor of normbreaking and
romantic relationships. We found no significant interaction between the two. Apparently,
early motor restlessness and pubertal timing are two independent predictors of later
normbreaking and romantic relationships with boys.
We also examined whether puberty accentuates motor restlessness over time (from age
10 to age 13), and whether such accentuation leads to problem behavior in mid-adolescence.
Motor restlessness changed from age 10 to age 13. This change is also a significant predictor
of later normbreaking and romantic relationships with boys, but the change did not seem to
interact with pubertal timing. Finally, pubertal maturation was found to be a unique predictor
of mid-adolescent normbreaking over and above the other terms in the model. This suggests
that early puberty does not accentuate individual differences behavior problems that existed
before puberty.
Some personality characteristics (thrill seeking, impulsivity, risk taking, and sensation
seeking) might be connected with an adventurous life style, including problem behavior an
dearly romantic relationships, and in fact, explain most of the interrelations we observed in
this paper (and the previous studies) and provide a reasonable rival hypothesis to the peer
socialization hypothesis.
The personality characteristics measures were significantly associated with problem
behavior, romantic relationships with boys and with peer involvement. Finally, all
characteristics except impulsivity correlated significantly with youth center attendance. They
did not, however, rival the peer socialization hypothesis. Thrill seeking and sensation seeking,
were not significantly associated with pubertal timing whereas the other two correlated with
pubertal timing on a low level. In a multiple regression analysis with delinquency as the
62
64
49
dependent variable, pubertal timing was a significant predictor even when all four personality
measures were entered in the first step.
In order to examine whether controlling for personality characteristics eliminates the
effects of the previously reported three-way interactions, we tested whether pubertal timing,
center attendance, and heterosexual (or peer) involvement predicts delinquency, and we
entered a broad measure of risk prone personality as a another main predictor. When the
measure of risk prone personality was entered as a predictor for delinquency along with
pubertal maturation, center attendance, and heterosexual involvement, pubertal maturation
remained a significant predictor. The slope of the interaction term Heterosexual involvement
x Center attendance x Puberty was almost identical to what it had been before risk prone
personality was entered. The findings were similar when heterosexual involvement was
substituted for peer interaction.
4.1.12 The conclusion of Study 4. We concluded that differences in mid-adolescent
problem behavior and romantic relationships with boys between early- and late-developing
girls cannot be accounted for by factors that are already present in late childhood or by
personality characteristics.
As a whole, Paper I lends strong support to the peer socialization hypothesis. The
alternative explanations that we tested did not provide an explanation for the link between
early pubertal timing and problem behavior.
4.2 Paper II
4.2.1 The research question of Paper II. The idea was to consider multiple
predictors simultaneously. I examined a biological factor (girls’ age at menarche), a
psychological factor (girls’ self-perceptions of maturity), and a social factor (early romantic
relationships with boys) simultaneously in an attempt to obtain a more comprehensive
understanding of the role played by pubertal timing in girls’ adjustment. For Paper I, we
assumed that early-developing girls feel more mature than their same-sex peers and that they
therefore affiliate with older peers. For Paper II, girls’ perceptions of their own maturity were
included in the reported analyses. It was hypothesized that early-developing girls who feel
mature and who have early romantic relationships with boys will have the most problematic
adjustment. I also argued that early-developing girls do not necessarily act more deviantly
simply because they feel mature. They may just as well act maturely and responsibly. Thus, it
was also hypothesized that early-developing girls who do not have early sexual relationships,
but who feel mature, will not show particularly high levels of problematic conduct.
4.2.2 The findings of Paper II. First, I examined the zero-order correlations between
pubertal timing, romantic relationships with boys, and self-perceptions of maturity on the one
hand, and the adjustment measures on the other. The results showed that early pubertal timing
was related to having problematic parent relations, poor school adjustment, a larger number of
peers, larger numbers of older and working peers, more normbreaking behavior, and greater
63
65
50
intoxication in adolescence. Early pubertal timing was associated with having older children,
low educational levels, and lower work positions in young adulthood.
Having romantic relationships by age 14-15 was related to having problematic parent
relations, poor school adjustment, a larger number of peers, larger numbers of older and
working peers, a higher frequency of meeting one’s peers, more normbreaking behavior, and
greater intoxication in adolescence. Having romantic relationships by age 14-15 was
associated with having older children, low educational levels, lower work positions, high
alcohol consumption, and criminality in young adulthood.
Perceiving oneself to be more mature than same-age peers in mid-adolescence was
related to having more peers, more older and working peers, and more normbreaking behavior
and drunkenness in mid-adolescence.
Second, I examined the eight subgroups of girls. They were based on early and late
pubertal timing, romantic relationships with boys, and self-perceptions of maturity. There
follows below a summary of how each group differed from the others (see also Table 2 in
Paper II). Only groups that differed significantly from the other groups are included.
• the Late Relationship-Only Group (late puberty, much relationships, low perceived
maturity): high alcohol consumption in young adulthood;
• the Late Mature-Only Group (late puberty, little relationships, high perceived
•
•
•
•
maturity): few adjustment difficulties in school and low frequency of intoxication in
adolescence and low number of children, high educational achievement, and high
work positions in young adulthood;
the Premature Group (early puberty, little relationships, low perceived maturity): low
level of normbreaking in adolescence;
the Early Relationship-Only Group (early puberty, much relationships, low perceived
maturity): poor school adjustment;
the Early Mature-Only Group (early puberty, little relationships, high perceived
maturity): the least problematic parent relations and low level of normbreaking in
adolescence;
the Early Precocious Group (early puberty, much relationships, high perceived
maturity): the most problematic parent relations, poor school adjustment, high number
of peers, older and working peers, frequent peer contacts, high level of normbreaking,
and intoxication in adolescence and high number of children; old first-borns, low
educational achievement, and low work positions in young adulthood.
In sum, there were significant differences between many of the subgroups. In line with
the hypothesis, the adjustment situations of early-developing girls were very different
depending on how they viewed themselves and whether they had early romantic relationships.
The Early Mature-Only group did not have any particular adjustment difficulties. By contrast,
the Early Precocious group had pronounced problems in many of the areas examined. Thus,
64
66
51
the combination of pubertal timing, perceived maturity, and romantic relationships seems to
be a fruitful predictor of female adjustment.
I also performed trajectory analyses using CFA. I examined the eight subgroups of girls
to see whether it was typical or antitypical for members of any of the groups to be mothers in
young adulthood (see Table 5 in Paper II). The trajectory analyses showed that it was typical
for members of the Early Precocious and the Early Relationship-Only groups to be mothers in
young adulthood. By contrast, it was not typical for members of the Early Premature and the
Early Mature-Only groups to have children in young adulthood. Furthermore, it was not
typical to have children in young adulthood among any of the trajectories involving latedeveloping females. By contrast, it was typical for the reference group (late puberty, little
experience of romantic relationships, and low perceived maturity) not to have children in
young adulthood. Thus, the trajectory analyses supported the previous findings by showing
that the life situations of early-developing females, even in adulthood, are radically different
depending on social and psychological contexts during adolescence.
Finally, I examined the life situations of women in midlife. There were few differences
between the sub-groups concerning midlife adjustment. There were, however, differences in
educational levels between the groups. The Late Mature-Only group showed the highest
educational achievements in midlife, and also had partners with the highest educational levels.
Early maturers who had much experience of early heterosexual relationships (the Precocious
and the Early Relationships-Only groups) had partners with the lowest educational
achievements.
4.2.2 The conclusion of Paper II. In conclusion, early pubertal timing was found to
be associated with adjustment difficulties in several areas, but when pubertal timing was
considered together with psychological and social factors, a more differentiated picture
emerged. In fact, the findings indicate that certain subgroups of early-developing females
might be among those who are the best adjusted.
4.3 Paper III
4.3.1 The research question of Paper III. The aim of Paper III was to examine the
short- and long-term implications of early female puberty concerning soma et psyche, since
there is little information about the long-term consequences of female pubertal timing.
4.3.2 The findings of Paper III. The results showed that at age 13 there was a
negative correlation between weight and height on the one hand and age at menarche on the
other, with the early-developing girls being heavier and taller than their peers. The degree of
pubertal maturation was also reflected by higher scores on the body mass index (BMI) among
the early-developing girls. The direction of difference in height between the early and the late
group was reversed by age 43, when the group of early-developing females was on average 3
cm shorter than the late-developing group. However, they were 6 kg heavier. Thus, BMI at
age 43 was 3 units higher among the women who had had an early menarche compared with
65
67
52
those with late menarche. In accordance with their higher BMI values, early maturers’
waistlines and the circumference of their hips were on average 6 cm wider than those of latedeveloping females. These differences were significant. At the age of 43, the early-developed
women were less physically fit and reported dieting more often than the other women. Earlydeveloping females were as happy with themselves, and their work and family, as laterdeveloping females, and they did not seem to differ in positive affect from the later-developed
women in midlife.
4.3.3 The conclusion of Paper III. In sum, this study indicates that early puberty has
long-term somatic effects in terms of weight status, body perception, and fitness, but no
psychological implications for women, at least not concerning life satisfaction. What is left to
find out is whether the implications of early puberty for weight status are similar for all earlydeveloping girls or whether there are subgroups of early maturers whose weight status might
be markedly different, and also whether childhood weight status explains the link between
early puberty and weight status.
4.4 Paper IV
4.4.1 The research question of Paper IV. The aim of Paper IV was to examine the
impact of pubertal timing on adult weight status, and the roles childhood weight status and
appetite might play in any such link.
4.4.2 The findings of Paper IV. The results indicate that body mass index (BMI)
tracks over time. BMI at age 8 correlated moderately with BMI at ages 25 and 37, and the
association between BMI at age 25 and age 37 was strong. Furthermore, age at menarche was
associated with BMI at all ages. Thus, girls with high BMI entered puberty sooner than
others; the earlier girls experienced menarche the higher was their BMI in adulthood.
We ran hierarchical regression analyses with adult BMI (ages 25 and 37) as the
dependent variable to examine whether pubertal timing had an independent predictive impact
on adult weight status over and above childhood weight status (see Table 3 in Paper IV).
Together, childhood BMI and age at menarche explained 31% of the variance in BMI at age
37, and they both emerged as significant predictors. BMI at age 8 emerged as the strongest
predictor. The results were similar for predicting BMI at age 25. Thus, age at menarche was
associated with adult weight status irrespective of childhood weight status.
Next, we tested whether appetite moderated the link between pubertal timing and adult
weight status. First, one year before menarche there was a trend for early-developing girls to
have larger appetites than late-developing girls. Moreover, appetite at age 8 and BMI at age 8
were not significantly correlated (p > .10). To test the moderating question (see Table 4 in
Paper IV), we entered age at menarche and appetite one year before age at menarche in the
first step of a hierarchical regression analysis predicting BMI at age 25. Together, age at
menarche and appetite explained 18% of the variance. Age at menarche emerged as the
strongest predictor. At a second step, we entered the interaction term (age at
66
68
53
menarche*appetite). Together, age at menarche, appetite, and the interaction term explained
26% of the variance, and all three emerged as significant predictors of BMI at age 25. We
performed the same procedure with BMI at age 37. The results were similar. Age at menarche
was not related to adult weight status for girls with small appetites, but was for those with
large appetites. Among girls with large appetites, early age at menarche was linked to high
adult weight status, but late age at menarche was not.
4.4.3 The conclusion of Paper IV. The findings suggest that pubertal timing is
associated with adult weight status irrespective of child weight status, but only under certain
circumstances. Appetite seems to moderate the link between pubertal timing and adult weight
status. Thus, although the findings support the idea that early pubertal timing is a risk factor
for high BMI, they also indicate that early-developing females will only have high BMI if
they have large appetites.
67
69
55
5. General discussion
5.1 The findings of the four papers and how they relate to previous
research
The biological, psychological, and social implications of girls’ puberty have long interested
social scientists. Today, that interest is more intense than ever. Just since 2005, 511 studies
have appeared in PsychINFO with the term puberty or pubertal in their abstracts (information
retrieved on March 3, 2008). Girls who enter puberty early are more likely than other girls to
develop internalizing and externalizing problems and to use substances (see reviews by
Alsaker, 1995, 1996; Buchanan et al., 1992; Graber, 2003; Stattin & Magnusson, 1990;
Susman et al., 2003; Susman & Rogol, 2004). Two additional major reviews have appeared
recently (Celio et al., 2006; Mendle et al., 2007). Thus, puberty is a hot topic and early
puberty, in particular, seems to be a risk factor for girls’ development.
Research on female pubertal timing has suffered from several major limitations. One is a
lack of good explanations for the link between early puberty and problem behavior. Another
is a lack of comprehensive research that includes biological, psychological, and social factors
when predicting any given outcome, and the general use of direct effect models rather than
interactive models when examining the possible implications of female pubertal timing for
behavior (see Graber, 2003). A third limitation is a lack of long-term follow-ups.
In a programmatic line of research, I have investigated the short- and long-term
developmental significance, concerning soma et psyche, of female pubertal timing. The
dissertation comprises four papers. The aim of each of those papers has been to overcome one
or several of the limitations from which the female pubertal timing literature has been
suffering.
5.1.1 Support for the peer socialization hypothesis. The peer socialization
hypothesis represents an attempt to give a comprehensive explanation for the link between
early puberty and problem behavior in adolescence. In brief, the peer socialization hypothesis
rests on: (a) ideas concerning selection and socialization from the literature on adolescent peer
affiliation (Dishion et al., 1994; Hartup, 1996; Kandel, 1978; 1986), and (b) the established
link between timing of puberty and initiation of sexual involvement (e.g., Aro & Taipale,
1987; Brown et al., 2005; Cavanagh, 2004; Flannery et al., 1993; Haynie, 2003; Jorm et al.,
2004; Rowe, Rodgers, & Meseck-Bushey, 1989; Silbereisen & Kracke, 1997; Simmons &
Blyth, 1987; Stattin & Magnusson, 1990; Udry & Billy, 1987). The hypothesis is that earlydeveloping girls select friends and romantic partners in the same way that most people tend to
select similar others to affiliate with. We tend to select others as friends and partners who in
important respects are similar to ourselves. In this case, early-developing girls will select and
be selected by peers who match them in terms of physical maturity. The peers who are similar
in maturity will typically be chronologically older, and in early adolescence this means that on
average they will be more involved in problem behaviors than the girls’ same-age peers. Once
69
71
56
early-developing girls form these peer affiliations, they will be socialized into the higher
levels of problem behaviors that those peers are already engaged in. The hypothesis is,
further, that early-developing girls will be more interested in romantic relationships than other
girls of their age, and, for this reason, some of the peers they affiliate with will be older boys,
who will on average be more involved in problem behavior than boys of their age. This will
intensify the socialization into problem behaviors.
For Paper I, we performed a series of studies designed to test the peer socialization
hypothesis at different contextual levels (cultural, school, and neighborhood). At a macro
level, we predicted and tested differences between contexts in which the peer socialization
effect should be more or less likely to appear. We did this by using samples from two
cultures. We also tested the peer socialization hypothesis at micro levels: first by comparing
peer affiliation in different everyday contexts like school and leisure, and then by comparing
girls who chose to spend their free time in a setting where problematic youths often
congregate with girls who spend their leisure time in other settings. Through this contextual
perspective on pubertal maturation, we attempted to overcome some of the limitations from
which the female pubertal timing literature has been suffering.
In the first study, we used samples from two different cultures – Sweden, in which the
attitudes toward teenage romance and sexuality are more tolerant, and Slovakia, in which the
attitudes are more restrictive. If contact with boys explains the link between early puberty and
problem behavior, then in Slovakia, where heterosexual contacts are more restricted, there
should be no link between pubertal timing and problem behavior. In Sweden, where
heterosexual contacts are more permissible, pubertal timing should be linked to problem
behavior. We found that in both cultures heterosexual contact was linked to problem behavior,
which is consistent with the peer-socialization hypothesis. As predicted, however, the link
between early puberty and problem behavior appeared in Sweden but not in Slovakia, thus
suggesting that norms concerning teenage romance and sexuality at a macro level can
facilitate or dampen the peer socialization effect. Thus, the results are consistent with the
predictions of the peer socialization hypothesis, and they show contextual amplification at the
level of cultural context.
In the second study, we turned to more proximal, everyday contexts – in-school versus
out-of-school contexts. According to the peer socialization hypothesis, early-developing girls
select friends and romantic partners who are similar to them, namely older peers. Previous
research, however, has found that early-developing girls do not differ from late-developing
girls with respect to proportion of older friends (Haynie, 2003). But this research used school
roster nomination strategies, which only allow participants to nominate peers at school as
friends. Our argument was that, at school, girls are restricted in affiliating with older peers
because of the age-graded nature of school activities; in contexts outside school, they are freer
to associate with older peers. This argument is in line with previous research showing that as
adolescent girls move from 6th to 10th grade, they affiliate more often with older boys, who
70
72
57
they typically meet outside school (Poulin & Pedersen, 2007). The prediction of the peer
socialization hypothesis is that early- and later-developing girls will have similar in-school
friends who do not differ in terms of chronological age or delinquency, but the out-of-school
friends of early-developing girls will be chronologically older and higher on delinquency than
the out-of-school friends of later-developing girls. We tested this hypothesis using data from a
whole-community study in which girls reported on the peers they engaged with in and out of
school. The results were consistent with our predictions. Early-developing girls had out-ofschool friends who were older and more delinquent than were the friends of the laterdeveloping girls. The friends met at school did not differ in age or delinquency between the
early- and later-developing girls. In these results there is evidence for the basic premise of the
peer socialization hypothesis – that early-developing girls affiliate with older peers, but only
in certain contexts.
In the third study of Paper I, we focused on a leisure context – neighborhood youth
centers – where delinquent youths are known to congregate (e.g., Mahoney et al., 2001).
According to the peer socialization hypothesis, if early-developing girls spend a lot of time in
contexts where delinquent boys hang out, they will be more like to be socialized into
delinquency than early-developing girls who do not spend time in those contexts. We
predicted and showed a 3-way interaction in which girls who (a) developed early, (b) attended
the youth centers, and (c) were romantically involved with boys showed the highest levels of
self-reported delinquency in adolescence and registered criminality in adulthood.
The findings that these types of free-time contexts facilitate the link between early
puberty and problem behavior is in line with previous research showing that early-developing
girls in adverse social contexts fare worse in terms of social adjustment than early-developing
girls in more affluent milieus (Biehl, Natsuaki, & Ge, 2007; Ge et al., 1996, 2002, 2006;
Obeidallah et al., 2004). Early-developing girls engage more in violent behavior and other
types of problem behavior in disadvantaged areas or in unsupportive families (Ge et al., 2002;
Obeidallah et al., 2004). Early-developing girls with substance abusing peers form more
favorable images of using substances than girls who develop later (Ge et al., 2006). Earlydeveloping girls who experience negative life events have higher levels of depressive
symptoms (Ge et al., 2001). Early-developing girls with mixed-sex friends show higher levels
of distress than other girls (Ge et al., 1996). Finally, early-developing youth with alcoholconsuming peers are likely to develop higher levels of alcohol use (Biehl et al., 2007). Thus,
the findings in this dissertation converge with previous research in showing that contextual
conditions play an important role in the link between pubertal timing and adjustment. In
addition, the present findings also highlight the mechanisms (i.e., peer socialization) that drive
the well-established associations between pubertal timing and problem behavior. The findings
of the first three studies provide empirical evidence for the ideas of peer socialization and
contextual amplification (see “Combining the peer socialization and contextual amplification
hypotheses” below), but it might be asked whether pre-existing characteristics of the girls
71
73
58
who choose to attend the youth centers and get involved with boys might explain their
problem behavior better than peer socialization does.
In the fourth study of Paper I, we answered the question raised above. Can differences in
problem behavior between early- and later-developing girls be explained by differences that
already existed before puberty? If they can, then early pubertal timing does not usher in new
behaviors, but transfers the differences that existed in childhood to adolescent life. We tested
this idea by examining individual differences in problem behavior at age 10 and personality
characteristics. The results showed that early problems and pubertal timing were independent
predictors of later problem behavior and romantic relationships with boys. Pubertal timing did
not seem to accentuate problem behavior for girls who were already problematic early on. We
concluded that the links between pubertal timing and mid-adolescent problem behavior, and
between pubertal timing and romantic relationships with boys, cannot be explained by
problems that existed before puberty or risk prone personality characteristics. We concluded
that the results presented in Paper I provide evidence for the peer socialization hypothesis.
The link between girls’ early puberty and problem behavior seems to be a function of normal
social processes in which youths affiliate with others who are similar to them. The peer
socialization hypothesis offers several advantages over other hypotheses that have been
forwarded to explain the link between early puberty and problem behavior as it offers a
specific explanation of the mechanisms linking pubertal timing and adjustment, and is able to
synthesize the existing literature.
5.1.2 Combining the peer socialization and the contextual amplification
hypotheses. The peer socialization hypothesis states the mechanisms linking early puberty
and problem behavior, or how this link is established. The contextual amplification
hypothesis, on the other hand, states the conditions under which the early puberty-problem
behavior link is most likely to occur. The conditions hypothesized most likely to strengthen
the early puberty-problem behavior link are adverse contexts, such as a poor neighborhood.
As with the peer socialization hypothesis, the contextual amplification hypothesis is well
supported by empirical findings (e.g., Biehl et al., 2007; Ge et al., 1996, 2002, 2006, in press;
Obeidallah et al., 2004). One of several explanations for contextual amplification is that girls
in troubled neighborhoods are more likely than other girls to meet and be influenced by
deviant peers (Ge et al., 2002). By combining the two hypotheses, it was possible for us to
specify some of the contextual conditions under which peer socialization should be more or
less likely to operate.
The findings in Paper I, specifically in studies 1 and 3, not only supported the peer
socialization hypothesis, but also the contextual amplification hypothesis (Ge et al., 2007). An
example of this was the finding in Study 3 that youth center attendance, which in some sense
could be called “adverse”, because the centers are visited by many deviant youth (Mahoney et
al., 2001), strengthened the association between early puberty and early romantic relationship,
and also that between early puberty and problem behavior.
72
74
59
The peer socialization and the contextual amplification hypotheses are probably best
viewed as complementary rather than competing. The peer socialization hypothesis states how
early puberty and problem behavior are linked. But for preventive and other purposes,
knowledge about when this process is most likely to occur is of central importance. This is
how the contextual amplification hypothesis enters the picture, since it states the contexts in
which this link should be the strongest. Combining the two ideas should give better
understanding of the developmental significance of female puberty, and a better chance to
predict when the link is likely to appear, thereby potentially enabling prevention of the
negative outcomes associated with early puberty.
The contextual amplification hypothesis does not specify under which societal or cultural
conditions the link between early pubertal timing and problematic adjustment is most likely to
occur, since it focuses more on micro and meso levels rather than the macro contextual level.
By contrast, in Paper I, we had hypotheses about macro contextual factors that might come
into play. In our predictions of the contexts in which we should expect a stronger link between
early puberty and problem behavior, we went beyond the idea of contextual amplification
hypothesis since we hypothesized and found support for the idea that if contact with boys
explains the link between early puberty and problem behavior, then in a cultural context,
where heterosexual contacts are more restricted, there will be no link between pubertal timing
and problem behavior. In a cultural context where heterosexual contacts are permissible,
pubertal timing will be linked to problem behavior. In conclusion, the peer socialization and
the contextual amplification hypotheses could be used simultaneously in order to get a better
understanding of the developmental significance of female puberty.
In Paper II, I examined the roles played by romantic relationships and subjective
perceptions of maturity in mid-adolescence in the link between female pubertal timing and
adjustment in adolescence and adulthood. I argued that a main effects approach to the study of
pubertal timing provides a limited means of fully understanding the role of pubertal timing in
girls’ adjustment. I aimed to overcome the problems of a lack of information about the longterm implications of female pubertal timing and of a lack of interactive models by
investigating pubertal timing in the context of psychological and social factors.
Aspects of maturity are typically linked to problem behavior in developmental theory
(e.g., Moffitt, 1993). In Paper I, we assumed that early-developing girls feel more mature than
their same-sex peers and that they therefore affiliate with older peers. This is an important
part of the peer socialization hypothesis linking early puberty and problem behavior.
However, it was never tested in that paper. For Paper II, a measure of girls’ subjective
perceptions of maturity was included in the analyses. Feeling older than one’s age and
subjective maturity have recently been linked to substance use and advanced sexual behavior
(Andershed et al., 2006; Arbeau et al., 2007). In accordance, I hypothesized that earlydeveloping girls who felt mature and who had early romantic relationships with boys (the
Early Precocious Group) should have high levels of problematic adjustment (see also Alsaker,
73
75
60
1995). But, maturity can mean many different things to young people (Tilton-Weaver et al.,
2001). Socially advanced behavior is certainly one, but responsible behavior, or genuine
maturity, is certainly another. Therefore, I argued, early-developing girls do not necessarily
act more deviantly simply because they feel mature. They might just as well act maturely and
responsibly. Thus, it was hypothesized that early-developing girls who do not have early
sexual relationships, but who feel mature, will not have particularly high levels of problematic
conduct (the Early Mature-Only Group). This is in contrast with some earlier developmental
theory (e.g., Moffitt, 1993).
The findings of Paper II were in line with this expectation. They showed that the Early
Precocious Group had high levels of problem behavior in mid-adolescence, in line with the
general idea that early puberty is linked to problem behavior (Alsaker, 1995, 1996; Celio et
al., 2006; Graber, 2003; Mendle et al., 2007; Stattin & Magnusson, 1990; Susman et al., 2003;
Susman & Rogol, 2004). In stark contrast to this was the finding that members of the Early
Mature-Only Group were among those girls who were in the least problematic adjustment
situations. This nuances the findings of Arbeau and colleagues (2007), who reported that
feeling older than one’s age is linked to problem behavior. It is also noteworthy that those
early-developing girls who were romantically involved early, but who did not feel more
mature than their peers (the Early Relationship-Only Group), did not stand out as problematic
either. Thus, arguing that feelings of maturity on behalf of early-developing girls are linked to
problem behavior seems to be a simplistic assumption. I conclude that pubertal timing or
feelings of maturity alone has limited predictive value. A biopsychosocial approach (Cairns,
1979; Bronfenbrenner & Ceci, 1994; Ford & Lerner, 1992; Magnusson, 1988; Magnusson &
Stattin, 1998), where pubertal timing is considered in the context of the developing girls’ view
of themselves, and in the context of romantic relationships, provides a much fuller picture of
the girl’s development. As such, the findings presented in Paper II build on and extend the
findings in Paper I.
An additional important piece of information is that there is a link between feeling older
than one’s age and the age of one’s current partner – with adolescent girls who feel older than
their age having older romantic partners (Arbeau et al., 2007). This might indicate that the
ages of the romantic partners of the Early Relationship-Only and the Early Precocious Groups
are different, with the Early Precocious Group having older romantic partners. Since older
adolescents have more problem behavior than younger (Sampson & Laub, 2003), this might
explain why the Early Precocious Group was more deviant than the Early RelationshipGroup. Further research is needed to discover if this is the case.
One of the aims of Paper III was to overcome the limitation of the literature concerning a
lack of information about the long-term implications of female pubertal timing. Very few
studies have examined what role pubertal timing plays in social adjustment in adulthood, and
even fewer of them have prospective longitudinal designs. For Paper III, we examined the
short- and long-term implications of female pubertal timing concerning soma et psyche. We
74
76
61
did this by using a sample of women who are now in their fifties, and who have been followed
in a longitudinal study from age 10. The results indicated that pubertal timing has implications
for females’ weight status in adolescence and in adulthood. In adulthood, early-developed
women were less physically fit and they reported dieting more often than other women. They
also had higher BMI and were shorter than other women. The latter finding is in line with
previous research (e.g., Freedman et al., 2003; Garn et al., 1986; Hulanicka et al., 2007; Must
et al., 2005; Wang et al., 2006). Early-developing females did not, however, differ from other
women in terms of perceived life satisfaction. Thus, we conclude that pubertal timing has
very long-term implications for women’s weight and physical status, and dieting. By contrast,
it seems to have little implications for adult women’s life satisfaction.
Paper III raises many questions. First, there is the traditional chicken and egg dilemma.
Does early puberty lead to high BMI or does high BMI lead to early puberty? Or, can the
tracking of BMI (Campbell et al., 2001; Sandhu et al., 2006) explain why early developing
girls have higher peri- and postpubertal BMI? Second, not all early developing girls have high
postpubertal BMI. Are there moderating factors affecting the strength of the relation between
early puberty and adult weight status?
In Paper IV, we attempted to answer these two questions. We did this by examining the
impact of pubertal timing on adult weight status, looking at the roles of childhood weight
status and appetite in this link. Appetite was included as a possible moderator. We used a
longitudinal sample of women, who have been followed from birth to adulthood. We tested
whether the link between pubertal timing and adult weight status still existed after statistically
controlling for childhood weight status. We also tested whether appetite moderated the link
between pubertal timing and adult weight status by employing a multiple regression analysis,
with interaction between pubertal timing and appetite as one of the predictors. The results
indicated that pubertal timing was associated with adult weight status irrespective of
childhood weight status. This finding is in contrast to earlier findings (Freedman et al., 2003;
Must et al., 2005), showing that the link between pubertal timing and adult weight status is
largely a result of childhood weight status affecting both pubertal timing (Adair & GordonLarsen, 2001; Davison et al., 2003; He & Karlberg, 2001; Juul et al. 2006; Kaplowitz et al.,
2001; Lee et al., 2007) and adult weight status (Campbell et al., 2001; Freedman et al., 2005;
Sandhu et al., 2006). Thus, this finding needs further exploration.
Previous research, including that presented in Paper III, has shown that early pubertal
timing predicts higher body mass in adolescence and adulthood (Adair & Gordon-Larsen;
Bini et al., 2000; Biro et al., 2001; Bratberg et al., 2007a; Freedman et al., 2003; Garn et al.,
1986; Hulanicka et al., 2007; Must et al., 2005; Villa et al., 2007; Wang et al., 2006). Yet, few
studies have examined whether there are conditions that moderate this link. One exception is
the study by Bratberg et al. (2007a), which showed that only early maturers with high waist
circumference in early adolescence had elevated levels of BMI in late adolescence. However,
because waist circumference was measured sometime between ages 12 and 16, puberty and
75
77
62
waist circumference were confounded in that study. In light of the previously reported
independent predictive impact of appetite on later body mass index (Lee & Song, 2007), we
examined whether appetite could be a moderator of the link between early puberty and adult
weight status. The results in Paper IV indicated that pubertal timing is not related to adult
BMI for girls with small appetites, but is for those with large appetites. Among girls with
large appetites, early puberty is linked to high adult BMI, but late puberty is not. Thus, we
conclude that early pubertal timing seems to predispose females to higher adult BMI,
irrespective of childhood BMI. But, their level of energy intake seems to make this outcome
more or less likely.
The mechanisms underlying this finding are unknown. Answers might lie in hormonal,
lifestyle, or psychological factors. Physical activity is one example of a lifestyle factor that
might help explain the finding. Energy intake and expenditure largely determine a person’s
BMI. Early-developing girls are less physically active than later-developing girls (Baker,
Birch, Trost, & Davison, 2007), and it might be that, given equal energy intake, their energy
balances are more likely to be in surplus than those of late-developing girls (see Garn et al.,
1986). But these are only speculations.
The findings in Paper IV might shed some new light on the literature showing a link
between early pubertal timing and somatic problems, such as breast cancer, and also
psychological problems, such as eating disorders and poor body perception. In conclusion, the
findings in Paper III are in accordance with previous research in suggesting that early pubertal
timing is linked to adult weight status. Paper IV builds on these findings in showing that
pubertal timing is an independent predictor of adult BMI. It also shows that it is only among
early-developing girls with high appetites that we find this link.
5.2 The studies as a whole – the biopsychosocial approach
The papers in this dissertation have all presented investigations of the developmental
significance of female puberty. They have concerned the biological, psychological, and social
implications of pubertal timing. Taken together, biological, psychological, and social factors
provide more comprehensive understanding of the role of pubertal timing in female
development, than if either one of those domains would have been examined in isolation.
Much of the research in this dissertation was guided by the biopsychosocial approach
(e.g. Cairns, 1979; Bronfenbrenner & Ceci, 1994; Ford & Lerner, 1992; Magnusson, 1988;
Magnusson & Stattin, 1998). According to the biopsychosocial model of development, there
is a dynamic interaction between biological, psychological, and social components, and these
interactions predict the development of any given outcome. As a whole, the papers indicate
that biological changes in puberty, and specifically the timing of those changes, play a
significant role in female development. However, all papers showed that knowledge about a
girl’s pubertal timing alone is not enough to make fruitful predictions about her development.
Rather, behavioral outcomes of pubertal timing depend on other vital circumstances.
76
78
63
A main effects approach can sometimes lead to erroneous conclusions. An informative
example comes from Swedish research on school performance. This research has shown that
immigrants perform worse in school, as measured by school grades (Swedish National
Agency for Education, 2004). This would be the conclusion if a main effects approach or a
direct effect model was used. However, if other characteristics of the school pupils are taken
into consideration, such as gender, a completely new picture emerges. Then, what researchers
have found is that immigrant girls perform better than boys born in and outside of Sweden
applies to (the Swedish National Agency for Education, 2004)! Thus, immigrant girls perform
better than Swedish boys. This authentic example clearly shows that a main effects approach
sometimes can be misleading and that interactive models, as those used in this dissertation,
are to be preferred.
Is puberty in general and early puberty in particular to be thought of as involving
pathological processes? “In many cases, authors have drawn conclusions about outcomes
from pubertal timing alone: …this risk factor of early maturation is significant and calls for
clinical and social intervention…” (Celio et al., 2006, p. 1269); “Detrimental psychological
outcomes associated with early pubertal timing in adolescent girls” (Mendle et al., 2007, p.
151); “Early pubertal timing associates with mental health problems in middle adolescence”
(Kaltiala-Heino, Marttunen, Rantanen, & Rimpelä, 2003, p. 1063); and,“Early puberty in girls
is associated with a startling number of psychopathologies and health problems” (Steingraber,
2007). I have a different viewpoint, which is based on the fact that there is little empirical
support for the common belief that maturing earlier per se as opposed to maturing later should
be particularly stressful and therefore result in various types of problems. In fact, research has
falsified the idea (Tschann et al., 1994; Wiesner & Ittel, 2002). I view early puberty as a nonpathological process. Rather than being a result of a pathological process, the link between
early puberty and problem behavior is a consequence of normal processes. The findings in
this dissertation clearly indicate that early maturation need not to be associated with
“detrimental psychological outcomes”. For instance, in Paper II, early-developing girls who
felt more mature than their classmates, and who were not particularly involved in romantic
relationships at an early age, were characterized by having the least problematic parent
relations and a low level of problem behavior in adolescence. Thus, the psychosocial
implications of pubertal timing seem to differ depending on other conditions in females’ lives.
The same phenomenon seems to be the case for bodily development. In this line of
inquiry, researchers have concluded that early pubertal timing is linked to higher body mass in
adulthood (e.g., Freedman et al., 2003; Garn et al., 1986; Hulanicka et al., 2007; Must et al.,
2005; Wang et al., 2006). This was also what we found in Paper III. The findings of Paper IV,
however, clearly revealed that the association between pubertal timing and adult weight status
only holds under certain conditions. According to the findings, this condition was a large
instead of a small appetite. As with the psychosocial implications, the bodily implications of
pubertal timing seem to differ depending on other conditions in females’ lives.
77
79
64
The majority of the papers in this dissertation have employed interactive models,
although admittedly they only comprise a small number of potential mediators and
moderators. Still, the findings demonstrate the importance of adopting a broad
biopsychosocial perspective and using interactive models in the study of the developmental
significance of female puberty. Future research, using large samples to avoid a lack of
statistical power and including even more potential moderators and mediators, should address
the issue of interactive approaches further.
In sum, the models we use for predicting the impact of female pubertal timing should not
only include measures of pubertal timing but also internal characteristics of developing girls
and variations in girls’ environments. This is the only way that we can fully understanding the
developmental significance of girls’ pubertal maturation and it will provide a favorable base
for developing effective preventive interventions.
5.3 Methodological limitations and strengths
This dissertation has certain limitations and strengths that should be noted. First, parts of the
papers in this dissertation were based on self-report questionnaires as the sole source of
information. This might be problematic as social desirability and other reporter biases might
come into play. In the majority of the studies, however, self-reports were used alongside other
sources of information. These sources of information were parental reports and physical
examinations conducted by medical staff.
Second, a related limitation is the use of a sole measure of pubertal timing (i.e., age at
menarche). This poses several problems. For instance, early-developing girls experience a
shorter period of time between the first appearance of secondary sexual characteristics and
menarche than later-developing girls (Pantsiotou et al., 2008). This means that, at pubertal
onset, the differences between girls with early and late menarche in secondary sexual
characteristics are larger than they are at the time of menarche. Using age at menarche as the
sole measure of pubertal timing might have had the effect of making the analyses more
conservative. Thus, for some of the research questions, it would perhaps have been better to
use a measure of age at pubertal onset instead of, or even better, in addition to, age at
menarche.
Third, some of the samples are rather old. Because things may have changed since the
data were collected, the extent to which the findings can be generalized to girls growing up
today is unclear. This problem, however, is inherent in any design involving long-term
follow-ups, and is not possible to avoid. Note that the findings in Paper I suggest that the role
of puberty, at least in adjustment, is similar now compared with what it was 30 years ago for
the issues of romantic relationships with boys and problem behavior.
Fourth, the vast majority of girls in all the samples were Caucasian. We do not know
whether the findings can be generalized to other ethnic groups. Concerning the peer
socialization idea, since there is a link between pubertal timing and sexual behavior and
78
80
65
problem behavior in other ethnic groups as well (Ge et al., 2002; Lam et al., 2002), it seems
likely that the same processes might also apply to other ethnic groups.
Fifth, socioeconomic status (SES) was not included in any of the analyses. In North
American samples, pubertal timing is linked to SES, with girls from lower socioeconomic
strata developing at an earlier age (Ellis & Essex, 2007). Given that SES is linked to problem
behavior, it could be a potential third variable explaining the link between early pubertal
timing and problem behavior. However, pubertal timing does not appear to be associated with
SES in Scandinavian samples (Bratberg, 2007). Therefore, the chance is small that the
findings in this dissertation could be explained by girls’ SES.
Sixth, all the studies were characterized by heteronormativity. This is a general limitation
in adolescent research. Adolescent romantic relationships constitute a major part of several of
the studies in this dissertation. Still, only heterosexual relationships were considered. We did
not explore the role of homosexual relationships in any of the papers.
Is it possible that early homosexual relationships mediate the link between early puberty
and problem behavior in the same manner as heterosexual relationships appear to? There
seems to be no association between age at menarche and sexual orientation among women
(Bogaert & Friesen, 2002). Moreover, since we expected heterosexual relationships to explain
the link between early puberty and problem behavior, based on previous findings showing
that boys are on average more deviant than girls (Lahey, Van Hulle, Waldman, Rodgers,
D’Onofrio, Pedlow, et al., 2006), homosexual relationships would not be expected to link
early puberty and problem behavior in the way heterosexual relationships seem to. Future
research is needed, however, to explore this issue.
Seventh, I have not distinguished between girls who experienced puberty early but within
the normal range and those with true precocious puberty. In all papers, girls with menarche
before age 10 were given the same score on the age-at-menarche measure. Not many girls
experienced menarche this early. The effect that this might have on the findings is that it
limits statistical power, and thus increases the chance of a Type II error. This means that the
statistical tests of this dissertation became more conservative as a result of this grouping of
girls.
Eighth, the scales that I and my colleagues have used in the papers have different
numbers of response options. When there are more response options, it is easier to detect a
correlation between two variables than when there are fewer response options. This is a
problem of restricted range.
This dissertation had a particular focus on the role of early puberty in female
development. Still, the role of puberty in female development might be curvilinear as well as
linear. Many of the analyses in this dissertation did not permit curvilinear interpretations of
the date. This is an issue for future research.
Despite these limitations, the papers in this dissertation have significant advantages.
These include, but are not limited to, the following: (1) the biopsychosocial approach to study
79
81
66
of the developmental significance of female puberty, (2) the use of several samples that differ
in terms of size, age of participants at time of data collections, the timing of the gathering of
information, and the country in which the data were collected, (3) the prospective longitudinal
research design, (4) the limited attrition rates, and (5) the close link between theory and
empirical testing. Perhaps the most important strength of this dissertation is the use of one
specific idea, the peer socialization idea, the systematic testing of this idea, and the support
these tests lent to the idea. It is my hope that other researchers in the pubertal timing area can
base their research on this part of this dissertation.
When interpreting the results, it is also important to note that the results apply to girls as a
group and not every single individual. Not all early-developing girls who have early sexual
relationships or who have a large appetite will end up with problem behaviors or a high adult
BMI, for instance. Furthermore, in this dissertation I have examined statistical significance,
not clinical relevance. These are two different things. Even though an association is
statistically significant, it might only be of modest clinical relevance. Thus, many earlydeveloping females thrive in adolescence and adulthood. The findings should be interpreted
with this in mind.
5.4 Implications for policy, prevention, and practice
The findings in this dissertation have several implications for future preventive efforts and
also for policy-making and practice. For professionals who work with children and teenagers,
the findings of this dissertation should provide valuable information. Healthcare personnel,
school teachers, social workers, parents, and others who come into contact with pubertal girls
should be informed about the role that pubertal timing plays in adolescent adjustment. As a
whole, the findings point to the need for selective preventive social interventions, targeting
early-developing girls and perhaps also their parents.
There is a clear difference between early pubertal timing and many other potential risk
factors1. By contrast with many other risk factors, there is not much to do about the early
maturation per se. The main factors that affect pubertal timing (e.g., genetic heritability) are
largely out of reach of preventive efforts. From a preventive perspective, perhaps the most
important and promising thing to note is that many, if not all, problems that are connected to
early puberty in girls develop several years after puberty has started. Here, there is room for
preventive efforts, whether they target early-developing girls in particular or whether they are
universal.
The findings have a particular bearing on school policies. In Sweden, sexuality education
in school is compulsory. Most of this education takes place during the 8th grade, when
students are around ages 14-15. The findings in this dissertation suggest that many girls, in
particular those who mature early, start having sex earlier than that. Early onset of intercourse
80
82
67
is a risk factor for many later problems (Cornelius, Clark, Reynolds, Kiriscia, & Tarter, 2007;
Magnusson, 2001). Furthermore, the earlier a girl has sex, the greater the chance that she does
not use any contraceptive (Wellings, Nanchahal, Macdowall, McManus, Erens, Mercer, et al.,
2001), hence increasing the risk of teenage pregnancy and sexually transmitted infections. If
we want sexuality education to count, and to matter for those who appear to be in need of it
the most, then school policies regarding such education need to change. One idea is that the
school nurse could complement ordinary teaching by addressing sexual issues earlier, during
regular visits in early adolescence, particularly focusing on early-developing girls.
What is the role of parents? Previous studies suggest that early-developing adolescents
have more conflicts with their parents than others (Steinberg, 1987). This dissertation shows
that it is the early-developing girls who feel more mature than their classmates, and who have
early romantic relationships, who have particularly high levels of problematic parent relations.
It is possible that good parenting may mitigate the negative effects that are associated with
early pubertal development in girls (Ge et al., 2002). Therefore, preventive interventions
could include parenting programs. Strengthening parent-child relations, increasing conflictresolution skills, and promoting family cohesion would be important components of any such
efforts. Research on whether this is the right way to tackle these issues needs to be conducted.
A few years ago, Swedish national radio broadcast a program about the secular trend in
pubertal maturation. The topics concerned were based on the fact that, over the last century,
girls have experienced menarche at earlier and earlier ages. There was a hotline that parents
could call. One parent who did was a mother who said that she was very worried because her
10 year-old daughter showed signs of impending puberty, and she asked whether this was to
be regarded as atypical or pathological.
This anecdote highlights the importance of being able to provide high-quality information
about the psychosocial meaning of early pubertal maturation to parents. Graber et al. (1997)
reasoned that “it is unlikely that parents are aware of the risks of early maturation for girls”
(p. 1774), and that they should receive comprehensive information about pubertal
development. I agree with the latter statement that parents should be informed about puberty.
I think it is important that parents are informed about what happens to their children as they
go through puberty, and that this information should include biological, as well as emotional
and social aspects of the pubertal transition. Such information would help parents assist their
children, and may also ease some of their own worries. To the contrary, I disagree with the
statement that parents are unaware of the risks associated with early puberty. I think it is very
likely, as the anecdote above suggests and previous research shows (O’Sullivan et al., 2000),
that parents believe that early puberty constitutes a risk in girls’ development. Therefore, it is
even more important to be able to give them accurate information. Just saying that early
maturation is a risk reflects an overly simplistic assumption. Early puberty in girls does not
11
In this dissertation, a risk factor is defined as a factor that is empirically linked to problem behavior. It does
not imply causality.
81
83
68
equal problems. As has been shown in this dissertation, it is only under certain circumstances
that early pubertal timing in girls is linked to adjustment difficulties. Therefore, parents
should be informed about the circumstances and conditions under which early puberty
constitutes a risk for problem development.
As girls go through puberty, giving parents and those who work with adolescent girls and
boys, accurate information about the potential risks associated with puberty, and in particular
with girls who develop early, may help those adults guide and assist adolescents as they
navigate their way through this eventful time of their lives. On the other hand, there is a great
risk that this type of information will be given to parents at the time when most adolescents
experience puberty. Then, however important the information may be, parents with earlydeveloping daughters will receive it too late for them to make the best use of it. Therefore, the
timing of the information provided to parents, teachers, and others, is of pivotal importance.
When girls experience puberty, they gain weight in terms of muscles and fat. This is how
it should be. A problem, however, might arise when girls gain too much weight (or too little).
The present dissertation suggests that early-developing females might be at higher risk of
developing overweight, regardless of their childhood BMI, than other girls. This, however,
seems only to hold for early maturers with large appetites. Thus, if overweight is to be
prevented, then targeting early-developing girls’ eating behaviors might be one measure that
can be taken. One has to bear in mind that I did not examine overweight as such, but rather
normal variations in BMI, and that the BMI of most early maturers may be considered as
normal. Nevertheless, the mean BMI of early-developing adult women, presented in Paper III,
was above the cut-off for overweight.
Finally, it is of crucial importance to promote research that develops and investigates
effective preventive efforts that can be directed at early-developing girls. This dissertation
suggests that is particularly those early-developed girls who initiate early romantic
relationships with boys and who feel mature who are at risk of adjustment difficulties and
who will have lower educational attainment in young adulthood. Therefore, specific efforts
should be developed, targeted, and evaluated with regards to these girls. Moreover, because
girls’ social networks seem to be a mechanism linking early pubertal maturation and
concurrent and future adjustment problems, preventive efforts should have elements of social
intervention.
5.5 Future research directions
Puberty constitutes one of the most radical changes in people’s lives, since it takes individuals
from reproductive immaturity to complete reproductive competence. Although the
psychological and social implications of puberty have long interested social scientists, many
questions remain to be answered.
This dissertation provides a theoretical framework for explaining the link between early
pubertal timing and problem behavior. Future research that builds on and tests this idea is
82
84
69
needed and encouraged. Although the findings in this dissertation supported the peer
socialization hypothesis, some of its premises were not tested. This should be investigated in
future studies.
Human behavior is contextual. It is difficult to image an acting individual without
thinking about the contextual conditions. Context is an important piece in the puzzle of the
link between puberty and adjustment. This is clearly shown in Paper I and Paper II and in
previous research (Biehl et al., 2007; Ge et al., 1996, 2002, 2006; Obeidallah et al., 2004).
However, this is an area that needs to be further developed. For instance, we will need to
achieve better understanding of the context in which early-developing girls spend their time.
Furthermore, an interesting question is how a girl’s own maturational standing affects the peer
group. Previous research, including the studies in this dissertation, has only focused on how
the peer group influences the developing girl.
More research should be devoted to testing and specifying the various assumptions of the
peer socialization hypothesis. A reason for this is that we need more detailed information in
order to know what advice to give to parents and professionals who work with young girls.
The role of parents is one aspect that is yet to be understood. One of the other unanswered
questions is how early-developing girls are brought into older peer networks. Are they the
ones who drive this process or is it older peers? There is still no information about the process
through which early maturing girls start to engage with older peers and boyfriends. A further
unanswered question is the role played by the early-developing girls’ best friends? Do earlydeveloping girls change best friends if they do not mature according to their own physical
development, or do they stay the same? This is an interesting issue in light of previous
findings showing that adolescent girls with romantic partners have better relationships with
their best friends (Kuttler & La Greca, 2004), which early-developing are more likely to have.
Methodology development is one area for future research. Specifically, the research field
should strive to achieve consensus concerning how puberty is to be operationalized and
measured. Indeed, some of the discrepancy in previous research may be due to
methodological issues (see Alsaker, 1996, 1995).
Future studies should attempt to incorporate several aspects of the pubertal changes, in
order to understand whether it is changes in hormone levels, the development of secondary
sexual characteristics, or any other changes during puberty, which are most important for the
short-term and long-term effects of pubertal timing. The measures should be of both objective
(e.g. physiological and anatomical measures) and subjective (e.g. self-reports of where in the
pubertal processes adolescent perceive that they are) nature. Much research on the
psychosocial aspects and meanings of puberty has used a sole indicator of puberty. This has
often been age at menarche. This is also true for this dissertation. In research on the
implications of pubertal timing on subsequent weight status and related issues, using several
measures of pubertal maturation might be particularly important as research has shown that
83
85
70
the correlation between age at menarche, which is a common measure of pubertal onset, and
the actual onset, has been decreasing and is now around .40 (Biro et al., 2006).
Also, understanding the connections between pubertal timing, adjustment, and cognitive
and neurological development should be an area for future inquiry. For instance, it would be
interesting to know whether the links between pubertal timing and adjustment are moderated
by the level of adolescents’ cognitive development.
While much research has focused on the developmental implications of female puberty,
male puberty has received relatively little attention. This is in contrast with many other
research areas. Thus, more research should be devoted to understanding the implications of
male puberty in general and its long-term effects in particular, since this is a field where
almost no research has been conducted.
5.5.1 Changing societies – does research keep up? Just as adolescence is a time
of change, the living conditions of adolescents change from one generation to the next.
Unfortunately, research questions and research designs in developmental research do not
always change accordingly. The development of information technology the last decade has
brought about a radical change in how adolescents live their lives. Mobile phones, computers,
and the Internet have entered adolescents’ lives and have strong impact on their social
relationships, school work, and leisure time. Today, the Internet is a very important medium
for adolescents (Rideout, Foehr, Roberts, & Brodie, 1999) and teenagers spend much of their
leisure time, and also school time, online. The vast majority of adolescents in Western
countries use it (Skoog, Stattin, & Kerr, in press); the Internet offers an array of activities and,
as a user, one can get information about virtually anything. This latter aspect is a source of
concern, for professionals, parents, and others alike. They pose questions like “What kinds of
information do adolescents come into contact with on the net?” “Can I find out which web
sites my son or daughter has visited?” and “Do I have the right to monitor my son’s or
daughter’s use of the Internet?”
Sexually explicit material is rife on the Internet, and a lot of the concerns that parents and
others have relate to it. Considering that most parents are unaware of what their children are
viewing online (Cameron, Salazar, Bernhardt, Burgess-Whitman, Wingood, & DiClemente,
2005), the concerns are scarcely surprising.
Information technology in general and adolescents’ use of the Internet in particular
provide a good example of a domain that has been subject to massive change lasting recent
years, and in which there is a lack of research and knowledge. To my knowledge, there is only
one study that has examined the role of pubertal timing in adolescents’ online activities
(Skoog et al., in press). In that study, we examined whether early, on-time, and latedeveloping boys differ in what they do online. First, we investigated boys’ access to
computers and the Internet. Second, we investigated what boys usually do online. Third, we
examined whom boys chat with on the Internet. We studied the puberty-Internet use link at
age 14, which according to past research, is around the time when pubertal timing varies most
84
86
71
among boys (Tanner, 1978). Virtually all the boys reported that they had access to a computer
at home and that they used the Internet. Early, on-time, and late-developing boys did not
differ in terms of looking for information for school, buying goods online, and most of the
other Internet activities. There was one key difference between the groups. Early maturers
reported downloading and viewing pornography significantly more often than the other boys.
The early maturers also differed from other boys in terms of watching and downloading films
and watching and downloading music, but these differences were not significant after
Bonferroni correction. Regardless of whether viewing pornography makes a positive, neutral,
or negative contribution to boys’ development, early maturers’ viewing and downloading of
pornographic material should be of relevance to parents and those working with youth. The
findings emphasize the need for further investigation of the impact of Internet pornography on
the development of adolescents.
In addition, that study highlights the need for research questions and research designs to
develop in concordance with the changing and developing living conditions of youths. Our
study (Skoog et al., in press) asked questions in a new area in which pubertal timing might
play a role, but it did not employ a novel or an innovative research design. This should be
explored further. In sum, the literature on pubertal timing and adjustment needs to make
advances, and investigate the living conditions of today’s, not yesterday’s, youth. As such, the
study calls for more timely reporting when exploring the role of pubertal timing in
adjustment.
5.6 Summary and final remarks
In a programmatic line of research, I have investigated the developmental significance,
concerning soma et psyche, of female puberty in this dissertation. Puberty is one of the most
radical transitions that people undergo. It signifies the change from sexual immaturity to
maturity and reproductive capability. It signals to the developing person, and also to others,
that a child is entering a new stage in life. Although puberty is a universal phenomenon that
all healthy people experience, individuals vary markedly with respect to the age at which
pubertal changes occur. A large body of literature has shown that when girls experience
puberty earlier than their peers they are at increased risk of later medical problems as well as
adjustment difficulties. The earlier girls mature, the more problems they seem to have.
However, there are few reasons to believe that direct effects models are appropriate for
studying the implications of pubertal timing (see Graber, 2003). Still, the majority of studies
have employed main effects approaches. The findings in this dissertation highlight the
importance of using interactive models, including psychological and social mediators or
moderators. This seems to hold for the implications of puberty for soma et psyche.
Concerning soma, the findings indicate that early puberty is linked to high body mass, not
only in childhood but also in adolescence and adulthood, and being less fit as an adult.
However, the link between early puberty and high adult weight status is only present under
85
87
72
certain circumstances; it is only when early-developing females have large appetites that they
have higher adult BMI. Thus, the variation in the timing of a normal biological process that
all healthy people undergo seems to have implications for future somatic status, but only
during certain circumstances.
Concerning psyche, the results suggest that early puberty seems to be a risk factor in
female development under certain circumstances. To be an early-developing girl who feels
mature and who has a socially advanced peer network, particularly early romantic
relationships with boys in mid-adolescence, seems to constitute a particular risk. This finding
supports the peer socialization hypothesis, according to which early-developing girls select
friends and romantic partners who are similar to them, in the same way that all people select
similar others with whom to affiliate. In this case, the peers who are similar in maturity will
be older, and in early adolescence this means that, on average, they will be more involved in
problem behaviors than the girls’ same-age peers. This is also true for the boyfriends of the
early maturers. Once early-developing girls form these affiliations, they will be socialized into
the higher levels of problem behaviors that those peers are already engaged in. Thus, the link
between girls’ early puberty and problem behavior seems to be a function of normal social
processes in which youths affiliate with others who are similar to them.
The main contributions of the research presented in this dissertation are that it: (1)
extended our understanding of the link between pubertal timing and girls’ problem behavior,
(2) provided a theoretical framework on which future research can be built, (3) documented
correlates concerning both soma et psyche of pubertal timing in adulthood, (4) highlighted the
importance of having a biopsychosocial approach, and (5) identified factors potentially
important for preventive interventions.
Eight hundred years before Christ, Hesiod said: “…for right timing is in all things the
most important factor”. In today’s language we would say “timing is everything”. Can this
common expression be applied to the developmental significance of female puberty? A large
body of research has shown that the timing of pubertal changes is important for adolescent
functioning. A recent publication has even been called “Puberty – Timing is everything!”
(Pinyerd & Zipf, 2005). But is timing really everything? The answer of this dissertation is no.
Timing seems to be an important piece in the puzzle, yes, but it is far from everything. It is
only by acknowledging that people are biological, psychological, and social beings, and that
their development is continuously affected by biological, psychological, and social factors,
that we can fully understand the developmental significance of female puberty. The fact that
psychological and social factors interact with pubertal timing in its role in human
development has important implications for prevention, since psychological and social
conditions and mechanisms can be the targets of preventive interventions.
86
88
73
5.7 Conclusions
• When, in relation to their peers, girls enter puberty seems to play a developmentally
significant biological, psychological, and social role in their adult and adolescent lives.
• Timing, however, does not seem to be everything with respect to the role of puberty in
female development, since pubertal timing seems to have different implications under
different conditions.
• The biopsychosocial model, which provided the theoretical foundation for the studies in this
dissertation, highlights the need to study puberty in the light of behavior and the social
context in order to get full understanding of its developmental significance.
• The peer socialization hypothesis, which says that early-developing girls move in peer
circles with more advanced social behavior, proved to provide a good explanation of the wellestablished link between pubertal timing and problem behavior. Thus, the link between girls’
early puberty and problem behavior seems to be a function of normal social processes in
which youths associate with others who are similar to them.
• Early pubertal puberty seems to predispose females to higher adult BMI, irrespective of
childhood BMI. But, females’ level of energy intake seems to make this outcome more or less
likely.
• Providing knowledge about the conditions that interact with pubertal timing with respect to
its role in development would be an important step in preventing the problems that have been
associated with early puberty among females.
87
89
75
References
Abassi, V. (1998). Growth and normal puberty. Pediatrics, 102, 507S-511S.
Adair, L. S., & Gordon-Larsen, P. (2001). Maturational timing and overweight prevalence in
US adolescent girls. American Journal of Public Health, 91, 642–644.
Alsaker, F. D. (1995). Timing of puberty and reactions to pubertal changes. In M. Rutter
(Ed.), Psychosocial disturbances in young people: Challenges for prevention (pp. 3782). Cambridge, UK: Cambridge University Press.
Alsaker, F. D. (1996). Annotation: The impact of puberty. Journal of Child Psychology and
Psychiatry, 37, 249-258.
Andershed, A-K., Johansson, T., & Pepler, D. (2006). The role of pubertal maturation in
adjustment among aggressive girls. Paper presented at the International Association for
Child and Adolescent Psychiatry and Allied Professions Congress, September, 2006,
Melbourne, Australia.
Andershed, H., Kerr, M., Stattin, H., & Levander, S. (2002). Psychopathic traits in nonreferred youths: A new assessment tool. In E. Blaauw, & L. Sheridan (Eds.),
Psychopaths: Current International Perspectives (pp. 131-158). The Hague: Elsevier.
Armour, S., & Haynie, D. L. (2007). Adolescent sexual debut and later delinquency. Journal
of Youth and Adolescence, 36, 141–152.
Arnett, J. J. (1999). Adolescent storm and stress, reconsidered. American Psychologist, 54,
317-326.
Aro, H., & Taipale, V. (1987). The impact of timing of puberty on psychosomatic symptoms
among fourteen- to sixteen-year-old Finnish girls. Child Development, 58, 261-268.
Baker, B. L., Birch, L. L., Trost, S. G., Davison, K. K. (2007). Advanced pubertal status at
age 11 and lower physical activity in adolescent girls. Journal of Pediatrics, 151, 488493.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social
psychological research: Conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, 51, 1173–1182.
Baumeister, L. M., Flores, E., & Marín, B. V. (1995). Sex information given to Latina
adolescents by parents. Health Education Research, 10, 233-239.
Bergman, L. R., & El-Khouri, B. M. (1998). SLEIPNER: A statistical package for
patternoriented analyses. Version II [Software manual]. Stockholm: Stockholm
University, Department of Psychology.
Biehl, M. C., Natsuaki, M. N., & Ge, X. (2007). The influence of pubertal timing on alcohol
use and heavy drinking trajectories. Journal of Youth and Adolescence, 36, 153-167.
Bini, V., Celi, F., Berioli, M. G., Bacosi, M. L., Stella, P., Giglio, P., et al. (2000). Body mass
index in children and adolescents according to age and pubertal stage. European
Journal Clinical Nutrition, 54, 214–218.
89
91
76
Biro, F. M., Huang, B., Crawford, P. B., Lucky, A. W., Striegel-Moore, R., Barton, B. A., et
al. (2006). Pubertal correlates in black and white girls. Journal of Pediatrics, 148, 234240.
Biro, F. M., McMahon, R. P., Striegel-Moore, R., Crawford, P. B., Obarzanek, E., Morrison,
J. A., et al. (2001). Impact of timing of pubertal maturation on growth in black and
white female adolescents: The national heart, lung, and blood institute growth and
health study. Journal of Pediatrics, 138, 636-643.
Blum, M., Harris, S. S., Must, A., Phillips, S. M., Rand, W. M., & Dawson-Hughes, B.
(2002). Household tobacco smoke exposure is negatively
premenopausal bone mass. Osteoporosis International, 13, 663-668.
associated
with
Bogaert, A. F., & Friesen, C. (2002). Sexual orientation and height, weight, and age of
puberty: New tests from a British national probability sample. Biological Psychology,
59, 135-145.
Bonat, S., Pathomvanish, A., Keil, M. F., Field, A. E., & Yanovski, J. A. (2002). Selfassessment of pubertal stage in overweight children. Pediatrics, 110, 743-747.
Bratberg, G. H. (2007). Pubertal timing – antecedent to risk or resilience? Epidemiological
studies on growth, maturation and health risk behaviours; The Young HUNT study,
Nord-Trondelag, Norway. Doctoral thesis, Norwegian University of Science and
Technology, Department of Public Health and General Practice.
Bratberg, G. H., Nilsen, T. I. L., Holmen, T. L., Vatten, L. J. (2007a). Early sexual
maturation, central adiposity and subsequent overweight in late adolescence. A fouryear low-up of 1605 adolescent Norwegian boys and girls: the Young HUNT study.
BMC Public Health, 7, 54.
Bratberg, G. H., Nilsen, T. I. L., Holmen, T. L., & Vatten, L. J. (2007b). Perceived pubertal
timing, pubertal status and the prevalence of alcohol drinking and cigarette smoking in
early and late adolescence: a population based study of 8950 Norwegian boys and girls.
Acta Paediatrica, 96, 292-295.
Bronfenbrenner, U. & Ceci, S. J. (1994). Nature-nurture in developmental perspective: A
bioecological theory. Psychological Review, 101, 568-586.
Brooks-Gunn, J., & Graber, J. A. (1994). Puberty as a biological and social event:
Implications for research on pharmacology. Journal of Adolescent Health, 15, 663-671.
Brooks-Gunn, J., Petersen, A. C., & Eichorn, D. (1985). The study of maturational timing
effects in adolescence. Journal of Youth and Adolescence, 14, 149-161.
Brown, J. D., Halpern, C. T., & L’Engle, K. L. (2005). Mass media as a sexual super peer for
early maturing girls. Journal of Adolescent Health, 36, 420-427.
Buchanan, C. M., Eccles, J. S., & Becker, J. B. (1992). Are adolescents the victims of raging
hormones? Evidence for activational effects of hormon on moods and behavior at
adolescence. Psychological Bulletin, 111, 62-107.
90
92
77
Buchanan, C. M., Eccles, J. S., Flanagan, C, Midgley, C, Feldlaufer, H., & Harold, R. D.
(1990). Parents' and teachers' beliefs about adolescents: Effects of sex and experience.
Journal of Youth & Adolescence, 19, 363-394.
Buhrmeister, D., & Furman, W. (1987). The development of companionship and intimacy.
Child Development, 58, 1101-1113.
Burt, S., McGue, M., DeMarte, J., Krueger, R. F., & Iacono, W. G. (2006). Timing of
menarche and the origins of conduct disorder. Archives of General Psychiatry, 63, 890896.
Cairns, R. B. (1979). Social development: The origins and plasticity of interchanges. San
Francisco: W. H. Freeman.
Cameron, K. A., Salazar, L. F., Bernhardt, J. M., Burgess-Whitman, N., Wingood, G. M., &
DiClemente, R. J. (2005). Adolescents' experience with sex on the Web: Results from
online focus groups. Journal of Adolescence, 28, 535-540.
Campbell, P. T., Katzmarzyk, P. T., Malina, R. M., Rao, D. C., Pérusse, L., & Bouchard, C.
(2001). Stability of Adiposity Phenotypes from Childhood and Adolescence into Young
Adulthood with Contribution of Parental Measures. Obesity Research, 9, 394-400.
Casey, V. A., Dwyer, J. T., Coleman, K. A., & Valadian, I. (1992). Body Mass Index from
Childhood to Middle Age: A 50-y Follow-Up. American Journal of Clinical Nutrition,
56, 14-18.
Caspi, A. (1995). Puberty and the gender organization of schools: How biology and social
context shape the adolescent experience. In L. J. Crockett & A. C. Crouter (Eds.),
Pathways through adolescence: Individual development in relation to social contexts.
The Penn State series on child & adolescent development (pp. 57-74). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Caspi, A., Lynam, D., Moffitt, T.E., & Silva, P. (1993). Unraveling girls' delinquency:
Biological, dispositional, and contextual contributions to adolescent misbehavior.
Developmental Psychology, 29, 19-30.
Caspi, A., & Moffitt, T. E. (1991). Individual differences are accentuated during periods of
social change: The sample case of girls at puberty. Journal of Personality and Social
Psychology, 61, 157-168.
Cavanagh, S. E. (2004). The sexual debut of girls in early adolescence: the intersection of
race, pubertal timing, and friendship group characteristics. Journal of Research on
Adolescence, 14, 285-312.
Cauffman, E., & Steinberg, L. (1996). Interactive effects of menarcheal status and dating on
dieting and disordered eating among adolescent girls. Developmental Psychology, 32,
631-635.
Celio, M., Karnik, N. S., & Steiner, H. (2006). Early maturation as a risk factor for aggression
and delinquency in adolescent girls: a review. International Journal of Clinical
Practice, 60, 1254-1262.
91
93
78
Compian, L., Gowen, L. K., & Hayward, C. (2004). Peripubertal girls’ romantic and platonic
involvement with boys: Associations with body image and depression symptoms.
Journal of Research on Adolescence, 14, 23-47.
Connolly, J. A., & Johnson, A. M. (1996). Adolescents' romantic relationships and the
structure and quality of their interpersonal ties. Personal Relationships, 3, 185-195.
Cornelius, J. R., Clark, D. B., Reynolds, M., Kirsci, M., & Tarter, R. (2007). Early age of first
sexual intercourse and affiliation with deviant peers predict development of SUD: A
prospective longitudinal study. Addictive Behaviors, 32, 850-854.
Costello, E. J., Sung, M., Worthman, C., & Angold, A. (2007). Pubertal maturation and the
development of alcohol use and abuse. Drug and Alcohol Dependence, 88S, S50-S59.
Crockett, L., & Petersen, A. C. (1987). Pubertal status and psychosocial development:
Findings from the Early Adolescence Study. In R. M. Lerner & T. T. Foch (Eds.),
Biological-psychosical interactions in early adolescence: A life-span perspective.
Hillsdale, NJ: Lawrence Erlbaum Associates.
Daniels, S. R., Khoury, P. R., & Morrison, J. A. (1997). The utility of body mass index as a
measure of body fatness in children and adolescents: differences by race and gender.
Pediatrics, 99, 804–807.
Darling, N., Dowdy, B. B., Van Horn, M. L., & Caldwell, L. L. (1999). Mixed-sex settings
and the perception of competence. Journal of Youth and Adolescence, 28, 461-480.
Davies, B. L. (1977). Attitudes towards school among early and late maturing adolescent
girls. Journal of Genetic Psychology, 86, 261-266.
Davison, K. K., Susman, E. J., & Birch, L. L. (2003). Percent body fat at age 5 predicts earlier
pubertal development among girls at age 9. Pediatrics, 111, 815-821.
Deardorff, J., Gonzales, N. A., & Christopher, F. S. (2005). Early puberty and adolescent
pregnancy: The influence of alcohol use. Pediatrics, 11, 1451–1456.
Dick, D. M., Rose, R. J., Pulkkinen, L. & Kaprio, J. (2001). Measuring puberty and
understanding its impact: A longitudinal study of adolescent twins. Journal of Youth
and Adolescence, 30, 385-400.
Dick, D. M., Rose, R. J., Viken, R. J., & Kaprio, J. (2000). Pubertal timing and substance use:
Associations between and within families across late adolescence. Developmental
Psychology, 36, 180-189.
Dishion, T. J., Patterson, G. R., & Griesler, P. C. (1994). Peer adaptations in the development
of antisocial behavior. In L. R. Huesmann (Ed.), Current perspectives on aggressive
behavior (pp. 61-95). New York: Plenum Press.
Dorn, L. D., Crockett, L. J., & Petersen, A. C. (1988). The relations of pubertal status to
intrapersonal changes in young adolescents. Journal of Early Adolescence, 8, 405-419.
Dorn, L. D., Dahl, R. E., Woodward, H. R., & Biro, F. (2006). Defining the boundaries of
early adolescence: A user's guide to assessing pubertal status and pubertal timing in
research with adolescents. Applied Developmental Science, 10, 30-56.
92
94
79
Dorn, L. D., Rotenstein, D. (2004). Early Puberty in Girls: The Case of Premature
Adrenarche. Women's Health Issues, 14, 177-183.
Dorn, L. D., Susman, E. J., & Ponirakis, A. (2003). Pubertal timing and adolescent adjustment
and behavior: conclusions vary by rater. Journal of Youth and Adolescence, 32, 157167.
Dubas, J. S. (2003). Pubertal Timing Effects on Adjustment During Adolescence and Young
Adulthood: Individual-, Peer-, and Parent-Amplification Effects. Paper presented at the
Society for Research on Adolescence conference, April, 2003, Tampa, Florida, USA.
Dubas, J. S., Graber, J. A., & Petersen, A. C. (1991). The effects of pubertal development on
achievement during adolescence. American Journal of Education, 99, 444–460.
Eichorn, D. H. (1975). Asynchronizations in adolescent development. In S. E. Dragastin & G.
H. Elder (Eds.), Adolescence in the life cycle: Psychological change and social context.
(xi, 324 pp.) Oxford, England: Hemisphere.
Elder, G. (1994). Time, Human Agency, and Social Change: Perspectives on the Life Course.
Social Psychology Quarterly, 57, 4-15.
Fergusson, D. M., Vitaro, F., Wanner, B., & Brendgen, M. (2007). Protective and
compensatory factors mitigating the influence of deviant peers on delinquent behaviours
during early adolescence. Journal of Adolescence, 30, 33-50.
Flannery, D. J., Rowe, D. C., & Gulley, B. L. (1993). Impact of pubertal status, timing, and
age on adolescent sexual experience and delinquency. Journal of Adolescent Research,
8, 21-40.
Ford, D. H., & Lerner, R. M. (1992). Developmental systems theory: An integrative approach.
Newbury Park, California: Sage Publications.
Freedman, D. S., Khan, L. K., Serdula, M. K., Dietz, W. H., Srinivasan, S. R., & Berenson, G.
S. (2003). The relation of menarcheal age to obesity in childhood and adulthood: the
Bogalusa Heart Study. BMC Pediatrics, 3, 3.
Freedman, D. S., Khan, L. K., Serdula, M. K., Dietz, W. H., Srinivasan, S. R., & Berenson, G.
S. (2005). The relation of menarcheal age to obesity in childhood and adulthood: the
Bogalusa Heart Study. Pediatrics, 115, 22-27.
Furman, W. (2002). The emerging field of adolescent romantic relationships. Current
directions in psychological science, 11, 177-180.
Furman, W., Brown, B. B., & Feiring, C. (1999). The development of romantic relationships
in adolescence. New York: Cambridge University Press.
Garn, S. M., LaVelle, M., Rosenberg, K. R., & Hawthorne, V. M. (1986). Maturational timing
as a factor in female fatness and obesity. American Journal of Clinical Nutrition, 43,
879-883.
Ge, X., Brody, G. H., Conger, R. D., & Simons, R. L. (2006). Pubertal transition and African
American children’s internalizing and externalizing symptoms. Journal of Youth and
Adolescence, 35, 531-540.
93
95
80
Ge, X., Brody, G. H., Conger, R. D., Simons, R. L., & Murry, V. M. (2002). Contextual
amplification of pubertal transition effects on deviant peer affiliation and externalizing
behavior among African American children. Developmental Psychology, 38, 42-54.
Ge, X., Conger, R. D., & Elder, G. H. (1996). Coming of age too early: Pubertal influences on
girls’ vulnerability to psychological distress. Child Development, 67, 3386-3400.
Ge, X., Conger, R. D., & Elder, G. H. (2001). Pubertal transition, stressful life events, and the
emergence of gender differences in adolescent depressive symptoms. Developmental
Psychology, 37, 404-417.
Ge, X., Elder, G. H., Regnerus, M., & Cox, C. (2001). Pubertal transitions, perceptions of
being overweight, and adolescents’ psychological maladjustment: Gender and ethnic
differences. Social Psychology Quarterly, 64, 363-375.
Ge, X., Jin, R., Natsuaki, M. N., Gibbons, F. X., Brody, G. H., Cutrona, C. E., et al. (2006).
Pubertal maturation and early substance use risks among African American children.
Psychology of Addictive Behaviors, 22, 404-414.
Ge, X., Natsuaki, M. N., Jin, R., & Biehl, M. (2007). A contextual amplification hypothesis:
Pubertal timing and girls’ problem behavior. Manuscript.
Goodson, P., Evans, A., & Edmundson, E. (1997). Female adolescents and onset of sexual
intercourse: A theory-based review of research from 1984 to 1994. Journal of
Adolescent Health, 21, 147-156.
Graber, J. S. (2003). Puberty in context. In C. Hayward (Ed.), Gender differences at puberty.
(pp.307-325). New York, NY, US: Cambridge University Press.
Graber, J. A., Brooks-Gunn, J., & Warren, M. (1995). The antecedents of menarcheal age:
Heredity, family environment, and stressful life events. Child Development, 66, 346359.
Graber, J. A., Brooks-Gunn, J. & Warren, M. P. (2006). Pubertal effects on adjustment in
girls: moving from demonstrating effects to identifying pathways. Journal of Youth and
Adolescence, 35, 413-423.
Graber, J. A., Lewinsohn, P. M., Seeley, J. R., & Brooks-Gunn, J. (1997). Is psychopathology
associated with the timing of pubertal development? Journal of the American Academy
of Child and Adolescent Psychiatry, 36, 1768-1776.
Graber, J. A., Seeley, J. R., Brooks-Gunn, J., & Lewinsohn, P. M. (2004). Is Pubertal Timing
Associated With Psychopathology in Young Adulthood? Journal of the American
Academy of Child and Adolescent Psychiatry, 43, 718-726.
Grumbach M. M., & Styne, D.M. (2003). Puberty ontogeny, neuroendocrinology, physiology,
and disorders. In Wilson, J. D. & Foster, P. W. (Eds.). Williams Textbook of
Endocrinology. (pp. 1139–1231). Philadelphia: W.B. Saunders.
Hall, G. S. (1904). Adolescence: Its psychology and its relation to physiology, anthropology,
sociology, sex, crime, religion, and education (Vols. I & II). Englewood Cliffs, NJ:
Prentice-Hall.
94
96
81
Hartup, W. W. (1996). The company they keep: Friendships and their developmental
significance. Child Development, 67, 1-13.
Harvey, S. M. & Spigner, C. (1995). Factors associated with sexual behavior among
adolescents: A multivariate analysis. Adolescence, 30, 253–264.
Haynie, D. L. (2003). Contexts of risk? Explaining the link between girls’ pubertal
development and their delinquency development. Social Forces, 82, 355-397.
Haynie, D. L., & Piquero, A. (2006). Pubertal Development and Victimization Risk:
Differences Across Gender. Journal of Research in Crime and Delinquency, 43, 3-35.
He, Q., & Karlberg, J. (2001). BMI in childhood and its association with height gain, timing
of puberty and final height. Pediatric Research, 49, 244-251.
Hockaday, C., Crase, S. J., Shelley, M. C., II, & Stockdale, D. F. (2000). A prospective study
of adolescent pregnancy. Journal of Adolescence, 23, 423-438.
Holmbeck, G., & Hill, J. (1988). Storm and stress beliefs about adolescence: Prevalence, selfreported antecedents, and effects of an undergraduate course. Journal of Youth &
Adolescence, 17, 285-306.
Hulanicka, B., Lipowicz, A., Kozie, S., & Kowalisko, A. (2007). Relationship between early
puberty and the risk of hypertension/overweight at age 50: evidence for a modified
Barker hypothesis among Polish youth. Economics and Human Biology, 5, 48-60.
Iannotti, R. J., Bush, P. J., & Weinfurt, K. P. (1996). Perceptions of friends' use of alcohol,
cigarettes, and marijuana among urban schoolchildren: A longitudinal analysis.
Addictive Behaviors, 21, 615-632.
Jones, H. E. (1938). The California adolescent growth study. Journal of Educational
Research, 31, 561–567.
Jones, M.C., & Mussen, P.H. (1958). Self-conceptions, motivations, and inter-personal
attitudes of early- and late-maturing girls. Child Development, 29, 491–501.
Jorm, A. F., Christensen, H., Rodgers, B., Jacomb, P. A. & Easteal, S. (2004). Association of
adverse childhood experiences, age of menarche and adult reproductive behavior: Does
the androgen receptor gene play a role? American Journal of Medical Genetics
(Neuropsychiatric Genetics), 125B, 105-111.
Juul, A., Teilmann, G., Scheike, T., Hertel, N. T., Holm, K., Laursen, E. M., et al. (2006).
Pubertal development in Danish children: comparison of recent European and US data.
International Journal of Andrology, 29, 247-255.
Kaltiala-Heino, R., Kosunen, E., & Rimpelä, M. (2003). Pubertal timing, sexual behaviour
and self-reported depression in middle adolescence. Journal of Adolescence, 26, 531545.
Kaltiala-Heino, R., Marttunen, M., Rantanen, P., & Rimpelä, M. (2003). Early puberty is
associated with mental health problems in middle adolescence. Social Science &
Medicine, 57, 1055-1064.
95
97
82
Kandel, D. B. (1978). Similarity in real-life adolescent friendship pairs. Journal of
Personality and Social Psychology, 36, 306-312.
Kandel, D. B. (1985). On processes of peer influences in adolescent drug use: A
developmental perspective. Advances in Alcohol and Substance Abuse, 4, 139-163.
Kandel, D. B. (1986). Processes of peer influences in adolescence. In R. K. Silbereisen, K.
Eyferth, & G. Rudinger (Eds.), Development as action in context: Problem behaviors
and normal youth development (pp. 203-227). Heidelberg: Springer-Verlag.
Kaplowitz, P. B., Slora, E. J., Wasserman, R. C., Pedlow, S. E., & Herman-Giddens, M. E.
(2001). Earlier onset of puberty in girls: relation to increased body mass index and race.
Pediatrics, 108, 347-353.
Karlberg, P., Engström, I., Lichtenstein, H., & Svennberg, I. (1968). The development of
children in a Swedish urban community. A prospective longitudinal study. III. Physical
growth during the first three years of life. Acta Paediatrica Scandinavica Suppl., 187,
48–66.
Keel, P. K., Fulkerson, J. A., & Leon, G. R. (1997). Disrodered eating precursors in pre- and
early adolescent girls and boys. Journal of Youth and Adolescence, 26, 203-216.
Kindblom, J. M., Lorentzon, M., Norjavaara, E., Lönn, L., Brandberg, J., Angelhed, J.-E., et
al. (2006). Pubertal timing is an independent predictor of central adiposity in young
adult males. Diabetes, 55, 3047-3052.
Kuttler, A. F., & La Greca, A. M. (2004). Adolescents' romantic relationships: Do they help
or hinder close friendships? Journal of Adolescence, 27, 395-414
La Greca, A.M., & Harrison, H. W. (2005). Adolescent peer relations, Friendships and
romantic relationships: Do they predict social anxiety and depression? Journal of
Clinical Child and Adolescent Psychology, 34, 49-61.
Lahey, B. B., Van Hulle, C. A., Waldman, I. D., Rodgers, J. L., D’Onofrio, B. M., Pedlow, S.,
et al. (2006). Testing descriptive hypotheses regarding sex differences in the
development of conduct problems and delinquency. Journal of Abnormal Child
Psychology, 34, 737-755.
Lam, T. H., Shi, H. J., Ho, L. M., Stewart, S. M., & Fan, S. (2002). Timing of pubertal
maturation and heterosexual behavior among Hong Kong Chinese adolescents. Archives
of Sexual Behavior, 31, 359-366.
Lanza, S. T., & Collins, L. M. (2002). Pubertal timing and the onset of substance use in
females during early adolescence. Prevention Science, 3, 69-82.
Larson, R., Clore, G., & Wood, G. (1999). The emotions of romantic relationships: Do they
wreak havoc on adolescents? In W. Furman, B. Brown, & C. Feiring (Eds.), The
development of romantic relationships in adolescence (pp. 19–49). Cambridge, UK:
Cambridge University Press.
Larson, R., & Richards, M. H. (1994). Divergent realities: The emotional lives of mothers,
fathers, and adolescents. New York: Basic Books.
96
98
83
Lee, J. M., Appugliese, D., Kaciroti, N., Corwyn, R. F., Bradley, R. H., & Lumeng, J. C.
(2007). Weight status in young girls and the onset of puberty. Pediatrics, 119, e624630.
Lee, K., & Song, Y-M. (2007) Parent-reported appetite of a child and the child’s weight status
over a 2-year period in Korean children. Journal of the American Dietetic Association,
107, 678-680.
Lemay, P. (1999). The statistical analysis of dynamics and complexity in psychology: a
configural approach. Doctoral thesis, l’Université de Lausanne, Faculty of the Social
and
Political
Sciences.
Retrieved
on
March
27,
2008
from
http://tecfa.unige.ch/~lemay/thesis/THX-Doctorat/THX-Doctorat.html
Lienert, G. A., & zur Oeveste, H. (1985). CFA as a Statistical Tool for Developmental
Research. Educational and Psychological Measurement, 45, 301-307.
Lindberg, S. M., Grabe, S., & Hyde, J. S. (2007). Gender, pubertal development, and peer
sexual harassment predict objectified body consciousness in early adolescence. Journal
of Research on Adolescence, 17, 723-742.
Lynne, S. D., Graber, J. A., Nichols, T. R., Brooks-Gunn, J., & Botvin, G. J. (2007). Links
between pubertal timing, peer influences, and externalizing behaviors among urban
students followed through middle school. Journal of Adolescent Health, 40, 181.e7–
181.e13.
Maggs, J.L., Almeida, D.M., & Galambos, N.L. (1995). Risky business: The paradoxical
meaning of problem behavior for young adolescents. Journal of Early Adolescence, 15,
339–357.
Magnusson, C. (2001). Adolescent girls’ sexual attitudes and opposite-sex relations in 1970
and in 1996. Journal of Adolescent Health, 28, 242-252.
Magnusson, D. (1988). Individual development from an interactional perspective. Hillsdale,
NJ: Erlbaum.
Magnusson, D., Dunér, A., & Zetterblom, G. (1975). Adjustment: A longitudinal study. New
York: Wiley.
Magnusson, D., & Stattin, H. (1998). Person-context interactions. In R. M. Lerner (Ed.),
Theoretical models of human development. Volume 1: Handbook of child psychology,
(5th ed., pp. 685-759). Editor-in-Chief: William Damon. New York: Wiley.
Magnusson, D., Stattin, H., & Allen, V. (1985). Biological maturation and social
development: A longitudinal study of some adjustment processes from mid-adolescence
to adulthood. Journal of Youth and Adolescence, 14, 267-283.
Magnusson, D., Stattin, H., & Allen, V. (1986). Differential maturation among girls and its
relation to social adjustment: A longitudinal perspective. In P. Baltes, D. Featherman, &
R. Lerner (Eds.), Life span development (Vol. 7). New York: Academic Press, pp. 134172.
97
99
84
Mahoney, J. L., Stattin, H., & Magnusson, D. (2001). Youth leisure activity participation and
individual adjustment: The Swedish youth recreation center. International Journal of
Behavioral Development, 509-520.
Manlove, J. (1997). Early motherhood in an intergenerational perspective: The experiences of
a British cohort. Journal of Marriage & the Family, 59, 263-279.
Markey, C. N., Markey, P. M., & Tinsley, B. J. (2003). Personality, puberty, and
preadolescent girls’ risky behaviors: Examining the predictive value of the five-factor
model of personality. Journal of Research in Personality, 37, 405-419.
Marshall, W. A., & Tanner, J. M. (1969). Variations in the pattern of pubertal changes
associated with adolescence in girls. Archives of Disease in Childhood, 44, 291–303.
Mendle, J., Turkheimer, E., & Emery, R. E. (2007). Detrimental psychological outcomes
associated with early pubertal timing in adolescent girls. Developmental Review, 27,
151-171.
Merriam-Webster online dictionary. (2007). Retrieved on March 27, 2008 from
http://www.m-w.com/dictionary/puberty.
Mezzich, A. C., Giancola, P. R., Lu, S. Y., Parks, S. M., Ratica, G. M., & Dunn, M. (1999).
Adolescent females with a substance use disorder: Affiliations with adult male sexual
partners. American Journal on the Addictions, 8, 190-200.
Miller, B. C., & Benson, B. (1999). Romantic and sexual relationships development during
adolescence. In W. Furman (Ed.) The Development of Romantic Relationships in
Adolescence (pp. 99–124), Cambridge: University Press.
Moore, K. A., Miller, B. C., Glei, D., & Morrison, D. R. (1995). Adolescent Sex,
Contraception, and Childbearing: A Review of Recent Research. Washington, D.C.:
Child Trends, Inc.
Mott, F. L., & Haurin, R. J. (1988). Linkages between sexual activity and alcohol and drug
use among American adolescents. Family Planning Perspectives, 20, 128-136.
Muir, A. (2006). Precocious Puberty. Pediatrics in Review, 27, 373-81.
Muris P., Meesters C., van de Blom W., & Mayer B. (2005). Biological, psychological, and
sociocultural correlates of body change strategies and eating problems in adolescent
boys and girls. Eating Behaviors, 6, 11–22.
Must, A., Naumova, E. N., Phillips, S. M., Blum, M., Dawson-Hughes, B., & Rand, M. W.
(2005). Childhood overweight and maturational timing in the development of adult
overweight and fatness: the Newton girls study and its follow-up. Pediatrics, 116, 620627.
Must, A., Phillips, S. M., Naumova, E. N., Blum, M., Harris, S., Dawson-Huges, B., et al.
(2002). Recall of early menstrual history and menarcheal body size: after 30 years, how
well do women remember? American Journal of Epidemiology, 155, 672-679.
98
100
85
Neemann, J., Hubbard, J., & Masten, A. S. (1995). The changing importance of romantic
relationship involvement to competence from late childhood to late adolescence.
Development and Psychopathology, 7, 727–750.
Nolen-Hoeksema, S. (1991). Responses to depression and their effects on the duration of
depressive episodes. Journal of Abnormal Psychology, 100, 569-582.
Obeidallah, D., Brennan, R. T., Brooks-Gunn, J., & Earls, F. (2004). Links between pubertal
timing and neighborhood contexts: Implications for girl’s violent behavior. Journal of
the American Academy of Child and Adolescent Psychiatry, 43, 1460-1468.
O’Sullivan, L. F., Heino, F. L., Meyer-Bahlburg, H. F. L., & Watkins, B. X. (2000). Social
cognitions associated with pubertal development in a sample of urban, low-income,
African-American and Latina girls and mothers. Journal of Adolescent Health, 27, 227235.
Otor, S. C. J., & Pandey, A. (1999). Adolescent transition to coitus and premarital
childbearing in Sudan: A biosocial context. Journal of Biosocial Science, 31, 361-374.
Pantsiotou, S., Papadimitriou, A., Douros, K., Priftis, K., Nicolaidou, P., & Fretzayas, A.
(2008). Maturational tempo differences in relations to the timing of the onset of puberty
in girls. Acta Paediatrica, 97, 217-220.
Parent, A-S., Teilmann, G., Juul, A., Skakkebaek, N. E., Toppari, J., & Bourguignon, J-P.
(2003). The timing of normal puberty and the age limits of sexual precocity: variations
around the world, secular trends, and changes after migration Endocrine Reviews, 24,
668-693.
Patton, G. C., Hibbert, M. E., Carlin, J., Shao, Q., Rosier, M., Caust, J., et al. (1996).
Menarche and the onset of depression and anxiety in Victoria, Australia. Journal of
Epidemiology and Community Health, 50, 661–666.
Patton, G. C., McMorris, B. J., Toumbourou, J. W., Hemphill, S. A., Donath, S., & Catalano,
R. F. (2004). Puberty and the Onset of Substance Use and Abuse. Pediatrics, 114, 300306.
Patton, G. C., & Viner, R. (2007). Pubertal transitions in health. Lancet, 369, 1130-1139.
Pedersen, W., Samuelsen, S. O., & Wichstrom, L. (2003). Intercourse debut age: Poor
resources, problem behavior, or romantic appeal? A population-based longitudinal
study. Journal of Sex Research, 40, 333 – 345.
Peskin, H. (1973). Influence of the developmental schedule of puberty on learning and ego
functioning. Journal of Youth and Adolescence, 2, 273-290.
Petersen, A. C., Crockett, L., Richards, M., & Boxer, A. (1988). A self-report measure of
pubertal status: Reliability, validity, and initial norms. Journal of Youth and
Adolescence, 17, 117-133.
Phinney, V. G., Jensen, L. C., Olsen, J. A., & Cundick, B. (1990). The relationship beween
early development and psychosexual behaviors in adolescent females. Adolescence, 25,
321-332.
99
101
86
Pietrobelli, A., Faith, M. S., Allison, D. B., Gallagher, D., Chiumello, G., & Heymsfield, S. B.
(1998). Body mass index as a measure of adiposity among children and adolescents: a
validation study. Journal of Pediatrics, 132, 204-210.
Pinyerd, B., & Zipf, W. B. (2005). Puberty-Timing is everything! Journal of Pediatric
Nursing, 20, 75-82.
Poulin, F., & Pedersen, S. (2007). Developmental changes in gender composition of
friendship networks in adolescent boys and girls. Developmental Psychology, 43, 14841496.
Rideout, V. J., Foehr, U. G., Roberts, D. F., & Brodie, M. (1999). Kids and media at the new
millenium: Executive summary. Menlo Park, CA: Kaiser Family Foundation (Nov.).
Rockhill, B., Moorman, P. G., & Newman, B. (1998). Age at menarche, time to regular
cycling, and breast cancer. Cancer Causes Control, 9, 447-453.
Rodriguez-Tomé, H., Bariaud, F., Cohen Zardi, M. F., Delmas, C., Jeanvoine, B., & Slylagyi,
P. (1993). The effects of pubertal changes on body image and relations with peers of the
opposite sex in adolescence. Journal of Adolescence, 16, 421-438.
Rogol, A. D., Roemmich, J. N., & Clark, P. A. (2002). Growth at puberty. Journal of
Adolescent Health, 31, 192–200.
Rowe, D. C., & Rodgers, J. L. (1994). A social contagion model of adolescent sexual
behavior: Explaining race differences. Social Biology, 41, 1-18.
Rowe, D. C., Rodgers, J. L., & Meseck-Bushey, S. (1989). An "epidemic" model of sexual
intercourse prevalences for Black and White adolescents. Social Biology, 36, 127-145.
Rutter, M., & Giller, H. (1984). Juvenile delinquency. Trends and perspectives. New York:
The Guilford Press.
Sampson, R. J., & Laub, J. H. (2003). Life-Course Desisters? Trajectories of Crime among
Delinquent Boys Followed to Age 70. Criminology, 41, 319-339.
Sandhu, J., Ben-Shlomo, Y., Cole, T. J., Holly, J., & Davey Smith, G. (2006). The impact of
childhood body mass index on timing of puberty, adult stature and obesity: a follow-up
study based on adolescent anthropometry recorded at Christ's Hospital (1936–1964).
International Journal of Obesity, 30, 14–22.
Sandler, D. P., Wilcox, A. J., & Horney, L. F. (1984). Age at menarche and subsequent
reproductive events. American Journal of Epidemiology, 119, 765-775.
Savin-Williams, R. C. & Berndt, T. J . (1990). ‘Friendship and peer relations’. In: Feldman,
S. S. and Elliott, G. R. (Eds), At the Threshold. The Developing Adolescent, pp.211-302,
Harvard University Press, Cambridge, MA.
Schlossberger, N. M., Turner, R. A., & Irwin, C. E. (1992). Validity of self-report of pubertal
maturation in early adolescents. Journal of Adolescent Health, 13, 109–113.
Schor, N. (1993). Abortion and adolescence: Relation between the menarche and sexual
activity. International Journal of Adolescent Medicine and Health, 6, 225-240.
100
102
87
Siegel, J. M., Yancey, A. K., Aneshensel, C. S., & Schuler, R. (1999). Body image, perceived
pubertal timing and adolescent mental health. Journal of Adolescent Health, 25, 155–
165.
Skoog, T., Stattin, H., & Kerr, M. (in press). The role of pubertal timing in what adolescent
boys do online. Journal of Research on Adolescence.
Silbereisen, R. K., & Kracke, B. (1993). Variation in maturational timing and adjustment in
adolescence. In S. Jackson & H. Rodriquez-Tomé (Eds.), The social worlds of
adolescence (pp. 67-94). East Sussex, UK: Erlbaum.
Silbereisen, R. K., Petersen, A. C., Albrecht, H. T., & Kracke, B. (1989). Maturational timing
and the development of problem behavior: Longitudinal studies in adolescence. Journal
of Early Adolescence, 9, 247-268.
Simon, A. E., Wardle, J., Jarvis, M. J., Steggles, N., & Cartwright, M. (2003) Examining the
relationship between pubertal stage, adolescent health behaviours and stress.
Psychological Medicine, 33, 1369-1379.
Simmons, R. G., & Blyth, D. A. (1987). Moving into adolescence: The impact of pubertal
change and school context. New York: De Gruyter.
Simmons, R. G., Blyth, D. A., & McKinney, K. L. (1983). The social and psychological
effects of puberty on white females. In J. Brooks-Gunn & A. C. Petersen (Eds.), Girls at
puberty. Biological and psychosocial perspectives (pp. 229-272). New York: Plenum
Press.
Simmons, R. G., Blyth, D. A., Van Cleave, E. F., & Bush, D. E. (1979). Entry into early
adolescence: The impact of school structure, puberty, and early dating on self-esteem.
American Sociological Review, 44, 948-967.
Simmons, R. G., Carlton-Ford, S. L., & Blyth, D. A. (1987). Predicting how a child will cope
with the transition to junior high school. In R. M. Lerner, & T. T. Foch (Eds.),
Biological-psychological interaction in early adolescence (pp. 325-375). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Smith, E. A., Udry, J. R. & Morris, N. M. (1985). Pubertal development and friends: A
biosocial explanation of adolescent sexual behavior. Journal of Health and Social
Behavior, 26, 183-192.
Smolak, L., Levine, M. P., & Gralen, S. (1993). The impact of puberty and dating on eating
problems among middle school girls. Journal of Youth and Adolescence, 22, 355–368.
Stattin, H., & Magnusson, D. (1990). Pubertal maturation in female development. Hillsdale,
NJ: Erlbaum.
Stattin, H., Magnusson, D., & Reichel, H. (1986). Criminality from Childhood to Aldulthood.
A Longitudinal Study of the Development of Criminal Behavior. The research program
individual development and adjustment. Report 63. Stockholm: Stockholms universitet,
Psykologiska institutionen.
101
103
88
Steinberg, L. (1987). Impact of puberty on family relations: Effects of pubertal status and
pubertal timing. Developmental Psychology, 23, 451-460.
Steinberg, L., & Morris, A. S. (2000). Adolescent development. Annual Review of
Psychology, 52, 83-110.
Steingraber, S. (2007). Falling age of puberty in U.S. girls: what we know, what we need to
know. A Breast Cancer Fund report. Retrieved December 4, 2007 from
http://www.breastcancerfund.org/site/pp.asp?c=kwKXLdPaE&b=3291891
Stice, E., Presnell, K., & Bearman, S. K. (2001). Relation of early menarche to depressionm
eating disorders, substance abuse, and comorbid psychopathology among adolescent
girls. Developmental Psychology, 37, 608-619.
Stone, C. P., & Barker, R. G. (1937). Aspects of personality and and intelligence in
postmenarcheal and premenarcheal girls of the same chronological age. Journal of
Comparative Psychology, 23, 439-445.
Stone, C. P., & Barker, R. G. (1939). The attitudes and interests of premenarcheal and
postmenarcheal girls. Journal of Genetic Psychology, 54, 27-71.
Susman, E. J., Dorn, L. D., & Schiefelbein, V. L. (2003). Puberty, sexuality, and health. In R.
M. Lerner & M. A. Easterbrooks (Eds.), Handbook of psychology: Developmental
psychology, vol. 6 (pp. 295-324). New York, NY: Wiley & Sons.
Susman, E. J., & Rogol, A. (2004). Puberty and psychological development. In R. M. Lerner
& L. Steinberg (Eds.), Handbook of Adolescent Psychology (pp. 15–44). New York:
Wiley.
Tanner, J. M. (1978). Foetus into man: physical growth from conception to maturity.
Cambridge, MA: Harvard University Press.
The Swedish National Agency for Education. (2004). Statistics on school marks. Retrieved
November 16, 2007 from http://www.skolverket.se/
Tilton-Weaver, L. C., Vitunski, E. T., & Galambos, N. L. (2001). Five images of maturity in
adolescence: What does "grown up" mean? Journal of Adolescence, 24, 143-158.
Tschann, J. M., Adler, N. E., Irwin Jr., C. E., Millstein, S. G., Turner, R. A., & Kegeles, S. M.
(1994). Initiation of substance use in early adolescence: The roles of pubertal timing and
emotional distress. Health Psychology, 13, 326-333.
Tremblay, L., & Frigon, J-Y. (2005). Precocious puberty in adolescent girls: a biomarker of
later psychosocial adjustment problems. Child Psychiatry & Human Development, 36,
73-94.
Turner, P. K., Runtz, M. G., & Galambos, N. L. (1999). Sexual abuse, pubertal timing, and
subjective age in adolescent girls: a research note. Journal of Reproductive and Infant
Psychology, 17, 111-118.
Udry, J. R. (1979). Age at menarche, at first intercourse, and at first pregnancy. Journal of
Biosocial Science, 11, 433-441.
102
104
89
Udry, J. (1988). Biological predispositions and social control in adolescent sexual behavior.
American Sociological Review, 53, 709-722.
Udry, J. R., & Billy, J. O. (1987). Initiation of coitus in early adolescence. American
Sociological Review, 52, 841-855.
van den Berg, S. M., Setiawan, A., Bartels, M., Polderman, T. J. C., van der Vaart, A. W., &
Boomsma, D. I. (2006). Individual differences in puberty onset in girls: Bayesian
estimation of heritabilities and genetic correlations. Behavior Genetics, 36, 261-270.
Veit, C. T., & Ware, J. E. Jr. (1983). The structure of psychological distress and well-being in
general populations. Journal of Consulting and Clinical Psychology, 51, 730-742.
Villa, I., Yngve, A., Poortvliet, E., Grjibovski, A., Liiv, K., Sjöström, M., et al. (2007).
Dietary intake among under-, normal, and overweight 9- and 15-year-old Estonian and
Swedish schoolchildren. Public Health Nutrition, 10, 311-322.
Wang, W., Zhao, L.-J., Liu, Y.-Z., Recker, R. R., & Deng, H.-W. (2006). Genetic and
environmental correlations between obesity phenotypes and age at menarche.
International Journal of Obesity, 30, 1595-1600.
Warren, M. P., & Brooks-Gunn, J. (1989). Mood and behavior at adolescence: Evidence for
hormonal factors. Journal of Clinical Endocrinology & Metabolism, 69, 77-83.
Waylen, A., & Wolke, D. (2004). Sex 'n' drugs 'n' rock 'n' roll: The meaning and social
consequences of pubertal timing. European Journal of Endocrinology, 151, 1-10.
Weichold, K., Pröhl S., Büttig S., & Silbereisen R. K. (2007). Pubertal timing and alcohol
use during early adolescence: investigating trajectories and mediational processes.
Paper presentation at the 13th European Conference on Developmental Psychology.
August, 2007, Jena, Germany.
Weichold, K., Silbereisen, R. K., & Schmitt-Rodermund, E. (2003). Short- and long-term
consequences of early versus late physical maturation in adolescents. In C. Hayward
(Ed.), Gender differences at puberty (pp. 241-276). Cambridge, MA: Cambridge
University Press.
Wellings, K., Nanchahal, K., Macdowall, W., McManus, S., Erens, B., Mercer, C. H., et al.
(2001). Sexual behaviour in Britain: Early heterosexual experience. Lancet, 358, 18431850.
Whiteman, S. D., McHale, S. M., & Crouter, A. (2007). Longitudinal changes in marital
relationships: The role of offspring's pubertal development. Journal of Marriage and
Family, 69, 1005-1020.
Wichstrom, L. (1999). The emergence of gender difference in depressed mood during
adolescence: The role of intensified gender socialization. Developmental Psychology,
35, 232-245.
Wichstrom, L. (2000). Psychological and behavioral factors unpredictive of disordered eating:
A prospective study of the general adolescent population in Norway. International
Journal of Eating Disorders, 28, 33-42.
103
105
90
Wichstrom, L. (2001). The impact of pubertal timing on adolescents’ alcohol use. Journal of
Research on Adolescence, 11, 131-150.
Wiesner, M., & Ittel, A. (2002). Relations of pubertal timing and depressive symptoms to
substance use in early adolescence. Journal of Early Adolescence, 22, 5-23.
Williams, J.M. & Currie, C.E. (2000). Self-esteem and physical development during early
adolescence: Pubertal timing and body image. Journal of Early Adolescence, 20, 129149.
Wilson, D. M., Killen, J. D., Hayward, C., Robinson, T. N., Hammer, L. D., Kraemer, H. C.,
et al. (1994). Timing and rate of sexual maturation and the onset of cigarette and alcohol
use among teenage girls. Archives of Pediatrics & Adolescent Medicine, 148, 789-795.
Woodward, L. J., Fergusson, D. M., & Horwood, L. J. (2001). Risk factors and life processes
associated with teenage pregnancy: Results from a prospective study from birth to 21
years. Journal of Marriage and the Family, 63, 1170-1184.
Zehr, J.L., Culbert, K.M., Sisk, C.L., & Klump, K.L. (2007). An association of early puberty
with disordered eating and anxiety in a population of undergraduate women and men.
Hormones and Behavior, 52, 427-435.
Zelnik, M., Kantner, J. F., & Ford, K. (1981). Sex and pregnancy in adolescence. Beverly
Hills: Sage.
Zimmer-Gembeck, M.J. (2002). The development of romantic relationships and adaptations
in the system of peer relationships. Journal of Adolescent Health, 31, 216-225.
Zimmer-Gembeck, M., Siebenbruner, J., & Collins, W. A. (2001). Diverse aspects of dating:
Associations with psychosocial functioning from early to middle adolescence. Journal
of Adolescence, 24, 313–336.
104
106
PAPER I
115
Klunserna Study
115
05-10-25, 10.22
Female
Pubertal
Timing
and Problem
Behavior
Female
Pubertal
Timing
and Problem
Behavior
1 1
Female Pubertal Timing and Problem Behavior:
Explaining the Mechanism at Different Levels of Social Contexts
Margaret Kerr
Therése Skoog
Håkan Stattin
Örebro University
Zdena Ruiselova
Institute of Experimental Psychology
Note: The three first authors are listed alphabetically as they have made equal contributions to the
study. The study was supported by a grant from The Swedish Council for Working Life and
Social Research.
PAPER I
Running Head: FEMALE PUBERTAL TIMING AND PROBLEM BEHAVIOR
2 Pubertal Timing and Problem BehaviorFemale Pubertal Timing and Problem Behavior 2
Female
Table of Contents
Chapter I: Introduction
Chapter II: The Broader, Cultural Context
Chapter III: School versus Free-Time Contexts
Chapter IV: Leisure Settings
Chapter V: Preexisting Differences
Chapter VI: General Discussion
References
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
We propose a contextual framework for explaining the link between pubertal
development and problem behavior in adolescent females. We explain under what
conditions we can expect early maturing girls to engage in unhealthy, dangerous, and
risky behavior, and under which conditions this is less likely. It is primarily under
conditions, at different contextual levels, that promote access to more advanced peer
groups, and particular opposite-sex relationships, that early pubertal maturation
should be linked with problem behavior. Under other conditions it is less likely.
We examine and confirm this hypothesis at the societal and everyday life levels,
comparing girls in two countries where the opportunities for accessing advanced peer
groups should vary considerably, and two populations of females who underwent
adolescence in one community but at different times. We then sought to qualify the
peer socialization hypothesis by differentiating between school and free-time friends.
In agreement with the theoretical notion, the link between early puberty and problem
behavior was found to be associated primarily with engaging with peers met in freetime settings, rather than with peers in the school setting. Third, we examined under
which everyday social conditions the association between early pubertal maturation
and problem behavior can be expected. In agreement with the theoretical argument,
early-maturing girls who frequently attended leisure settings that contained access to
older peers and boys developed social problem behavior over time, whereas this was
less true for other early and later maturing girls. Overall, we conclude that the
theoretical model proposed satisfactorily explains the link between pubertal timing
and problem behavior at different levels of social contexts.
3
PAPER I
Abstract
3
4Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
Chapter 1: Introduction
There are few areas in research on adolescent development that have received so
much interest from researchers from various disciplines as puberty and individual
differences in the timing of its occurrence (for a review, see Susman, Dorn, &
Schiefelbein, 2003). Radical bodily changes and hormonal changes occur during a
short time span (Buchanan, Eccles, & Becker, 1992) and they seem to have
substantial consequences for behavior and attitudes. With the changes that occur in
height, weight, and secondary sexual characteristics, puberty signals that the young
person is leaving childhood and is ready for a new phase in life. Adults and youths
associate physical maturation with social maturation and being “grown up” (Johnson
& Collins, 1988; Tilton-Weaver, Vitunski, & Galambos, 2001). Paradoxically,
though, girls who go through puberty early do not typically tend to show adult-like,
responsible behavior. Instead, they are overrepresented among those who show
problem behaviors such as delinquency.
Well-documented negative correlations exist between girls’ pubertal timing and
various measures of problem behavior. Well over one hundred empirical studies have
been devoted to this issue the last 20 years, some of which have made significant
contributions to understanding of the role of pubertal timing in adolescent girls’
problem behaviors. (e.g., Aro & Taipale, 1987; Burt, McGue, DeMarte, Krueger, &
Iacono, 2006; Caspi, 1995; Caspi & Moffitt, 1991; Caspi, Lynam, Moffitt, & Silva,
1993; Dick, Rose, Viken, & Kaprio, 2000; Ge, Brody, Conger, & Simons, 2006; Ge,
Brody, Conger, Simons, & Murry, 2002; Ge, Conger, & Elder, 1996; Ge, Jin,
Natsuaki, Gibbons, Brody, Cutrona, et al., 2006; Graber, Lewinsohn, Seeley, &
Brooks-Gunn, 1997; Graber, Seeley, Brooks-Gunn, & Lewinsohn, 2004; Haynie,
2003; Lanza & Collins, 2002; Lynne, Graber, Nichols, Brooks-Gunn, & Botvin,
2007; Martin, Kelly, Rayens, Brogli, Brenzel, Smith, et al., 2002; McMaster,
Connolly, Pepler, & Craig, 2002; Obeidallah, Brennan, Brooks-Gunn, & Earls, 2004;
Patton, McMorris, Toumbourou, Hemphill, Donath, & Catalano, 2004; Silbereisen &
Kracke, 1993; Silbereisen & Kracke, 1997; Simmons & Blyth, 1987; Simmons, Blyth,
& McKinney, 1983; Simmons, Blyth, Van Cleave, & Bush, 1979; Simmons, CarltonFord, & Blyth, 1987; Stattin & Magnusson, 1990; Susman, Dockray, Schiefelbein,
Herwehe, Heaton, & Dorn, 2007; Tschann, Adler, Irwin, Millstein, Turner, &
Kegeles, 1994; Wichstrom, 2001; Wiesner & Ittel, 2002; Wilson, Killen, Hayward,
Robinson, Hammer, Kraemer, et al., 1994). Early maturing girls have higher rates of
experimental and regular tobacco use, externalizing problems, norm violations,
delinquency, drug and alcohol use, sensation seeking, antisocial personality, and
psychopathology, aggression, school problems, including non-attendance and lack of
school motivation, engagement with deviant peers, sexual harassment of other-sex
peers, and peer support for normbreaking (for reviews, see Alsaker, 1995, 1996;
Buchanan et al., 1992; Celio, Karnik, & Steiner, 2006; Mendle, Turkheimer, &
Emery, 2007; Susman et al., 2003; Susman & Rogol, 2004). Similar findings have
been reported in studies of girls with extremely early, or precocious, puberty (Comite,
Pescovitz, Sonis, Hench, McNemar, Klein, et al., 1987; Ehrhardt, Meyer-Bahlburg,
Bell, Cohen, Healy, Stiel, et al., 1984; Sonis, Comite, Blue, Pescovitz, Rahn, Hench,
et al., 1985). These results appear in North American and European samples, for
Anglo-American and African American adolescents, and in cross-sectional and
longitudinal studies. They are quite consistent and robust over different types of
adolescent populations.
In addition to normbreaking, early puberty in girls has been linked to clinical
levels of behavior problems. For instance, in an epidemiological study involving a
4
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
5
Explanations for the link between early puberty
and problem behavior
It is not readily apparent why girls who enter puberty early some years later appear
more normbreaking than girls who mature on time or late. Also, one would ideally
like to have an explanation that accounts for the few findings in the literature that
pubertal timing has no association with problem behavior. A number of hypotheses
have been advanced to explain why early maturing girls might have problems in
adjustment that lead them into normbreaking or antisocial behavior.
Third Variables and Pre-existing Problems. On hypothesis is that
problems existed before early puberty, but are not caused by it. For instance, there
might be genetically based characteristics or family factors that underlie both early
puberty and problem behavior. If so, the link between pubertal timing and problem
behavior might be spurious, and if so, no further explanations are needed. Pubertal
timing is substantially governed by genetic factors. Analyses of data from two large
national studies of adolescents, reported that half of the variance in pubertal timing
was attributable to genetic factors and the rest to nonshared environmental factors
(Ge, Natsuaki, Neiderhiser, & Reiss, 2007). The possibility that genetic factors
underlie both early pubertal timing and problem behavior was investigated in a birth
cohort of Finnish twins using smoking and drinking as the problem behaviors of
interest (Dick et al., 2000). In between-family analyses, early development was linked
PAPER I
large population of 16-year-old adolescents, the early maturing girls had higher rates
of substance use, conduct and oppositional disorders, and hyperactivity than on-time
or late maturing girls (Graber et al., 1997). When followed up at age 24, the early
maturing girls had higher lifetime prevalences of disruptive behaviors and antisocial
personality than did the on-time and late developing girls, indicating that latermaturing girls did not catch up in terms of problem behavior (Graber et al., 2004).
Similarly, early maturing girls in an outpatient clinical sample had higher levels of
school maladjustment, parental conflicts, and asocial behavior than the other girls in
the sample (Frisk, 1968; Frisk, Tenhunen, Widholm, and Hartling, 1966). In sum, the
link between problem behaviors and girls’ early development appears for girls whose
puberty is on the early end of the normal range and for those with clinically early, or
precocious, puberty; it appears in clinical and nonclinical samples; and it appears for
clinical levels of problem behaviors as well as normbreaking.
It is difficult to escape the conclusion that girls who go through puberty early are
more likely than their peers to get involved in various forms of problem behavior.
Despite this, there are a few studies that do not agree. Some have failed to find
relations between early development and problem behavior (Crockett & Peterson,
1987; Dubas, 2002; Duncan, Ritter, Dornbusch, Gross, & Carlsmith, 1985; Graber &
Brooks-Gunn, 1996; Lenerz, Kucher, East, Lerner, & Lerner, 1987; Mezzich, Tarter,
Hsieh, & Fuhrman, 1992; Gold & Petronio, 1980; Wadsworth, 1979). Others have
found pubertal timing to be related to some problem behaviors but not others (e.g.,
Martin et al., 2002; Graber et al., 1997). Some studies have failed to link early
development with poor school adjustment and achievement (Duke, Carlsmith,
Jennings, Martin, Dornbusch, Gross, et al., 1982; Duncan et al., 1985), and older
British studies have reported the opposite—that early maturing girls were better
adjusted and achieved better at school (Douglas, 1964; Douglas & Ross, 1964;
Poppleton, 1968; see also Abernethy, 1925). Thus, although the overwhelming
majority of studies in the literature link early development with problem behavior,
there are some that do not.
5
6Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
to more and earlier smoking and drinking, as expected. Notably, however, the results
were similar in within-family analyses comparing early maturing girls with twin
sisters who matured at least 2 years later. Early maturing girls had more and earlier
smoking and drinking than their later-maturing twins. This suggests that, at least for
smoking and alcohol drinking, genetic and family factors do not explain the link
between early development and problem behaviors. Similar conclusions, that there is
no common genetic influence for pubertal timing and problem behavior have been
drawn recently by others (Burt et al., 2006; Jorm, Christensen, Rodgers, Jacomb, &
Easteal, 2004). Nor have strong links been found between menarcheal age and
childhood conduct problems (Burt et al., 2006; Caspi & Moffitt, 1991; Graber,
Brooks-Gunn, & Warren, 1995; Moffitt, Caspi, Belsky, & Silva, 1992). In conclusion,
the current evidence does not seem to support ideas about common genetic influence
or pre-existing social adjustment problems among early maturing girls.
Another explanation for the link between early pubertal timing and normbreaking
behavior rests on the idea that differences between people will be magnified and
accentuated during periods of discontinuity, and puberty is such a period (Caspi &
Moffitt, 1991). This argument, known as the accentuation hypothesis, is that puberty
should enlarge individual differences that existed before puberty. The accentuation
hypothesis was examined in the Dunedin longitudinal study (Caspi & Moffitt, 1991).
The early maturing girls increased in problem behavior more than did on-time and late
maturing girls, but this increase was highest for the early maturing girls who had high
levels of problem behavior before puberty. The accentuation hypothesis does not
explain why early maturing girls go into problem behaviors more than late maturing
girls, but it suggests that the acceleration in problem behavior is dependent on
conditions that were at hand already much earlier. So far, empirical confirmation
exists only for one longitudinal cohort. The idea should be tested in different
longitudinal studies. The advantage of the accentuation hypothesis is that it links what
has been found for early developing girls to a more universal developmental principle.
The disadvantage is that it does not explain why this process should work more for
early developers than for on-time or late developers.
Emotional Distress Due to Bad Timing. There are two other
explanations that are so closely related that in many ways it is difficult to differentiate
them. The off-time hypothesis (Brooks-Gunn et al., 1985, see also Brim & Ryff, 1980;
Neugarten, 1969; Neugarten, Moore & Lowe, 1965; Eichorn, 1975; Petersen &
Taylor, 1980), sometimes called the maturational deviance hypothesis (Alsaker,
1995b), states that reaching important milestones in development earlier or later than
expected will be experienced as stressful and problematic. The early maturing girl is
emotionally unprepared, behaviorally unskilled, and cognitively too immature to
engage in the advanced behaviors commensurate with her physical maturity. In
addition, she has few same-age peers from which to gain information and support
during the time of the transition. Later-developing girls should be similarly out of
synch with their peers. Thus, early-developing girls should have problems in early
adolescence, and later-developing girls should have problems in later adolescence.
Although both the early- and the later-developed girls should differ from those who
are on time, one should perhaps not expect the same consequences for them.
Early-developing girls are thought to have trouble coping because they are
emotionally immature and unprepared for the changes of puberty (Graber, Petersen, &
Brooks-Gunn, 1996). But the “off-time” idea does not readily explain why early
matured girls are more likely than other girls to engage in norm-breaking behaviors.
One might think the opposite. Girls who develop early will come to see themselves as
6
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
7
PAPER I
different from others, and more mature, and for this reason adopt more adult-like
behavior than their age-mates. Also, with the parallel argument of being out of sync
with one’s peers, the idea that the later-developing girls should be higher on norm
breaking in later adolescence could be applied. But, there is no empirical evidence for
elevated levels of norm violating or substance use for late maturing girls during the
time when they are going through puberty (cf. Caspi & Moffitt, 1991; Ge, Conger, et
al., 1996; Tschann et al., 1994). Thus, being off time, alone, does not seem to be
linked to problem behavior. Thus, some expectations that follow from the off-time
hypothesis are not supported empirically.
Closely related to the off-time hypothesis is the early timing hypothesis (BrooksGunn, Petersen, & Eichorn, 1985; Graber, 2003) or the stage termination hypothesis
(Petersen & Taylor, 1980). The idea is similar to the off-time hypothesis except that it
focuses on being early rather than being off-time, per se. In addition, some scholars
have brought up the idea that girls who mature early do not get to experience much of
a non-stressful late childhood, because it is interrupted by the onset of puberty
(Peskin, 1973). Others have argued that early developing girls are forced to adopt
adult-like behavior earlier than others, which they perceive as stressful (Brooks-Gunn
et al., 1985) or that they have limited opportunities to engage in adult-like behaviors,
which they desire to, so they turn to normbreaking as a way of expressing adult-like
behavior (Moffitt, 1993). According to the early timing hypothesis, to be early
maturing is to experience pubertal maturation in isolation. Furthermore, pubertal
development is associated with more body fat, and this makes early developers
dissatisfied with their bodies more than later developing girls (Crockett & Petersen,
1987; Simmons et al., 1979). Thus, this hypothesis deals with timing issues that are
unique to early-developers.
Both the off-time and early timing hypotheses suggest that being out of synch
with peers creates internal distress, which in turn, spurs girls to engage in
normbreaking behavior. They put much emphasis on the emotional stress that comes
with developing out of sync with one’s peers (Brooks-Gunn & Graber, 1994; Ge,
Conger, & Elder, 2001; Steinberg & Morris, 2000). The questions, then, are whether
early maturing girls show more distress and internal problems than on-time and late
maturing girls and whether internal distress leads to problem behavior.
The literature offers different views about the link between early puberty and
internal distress. For instance, in the 1990s, one major review concluded that pubertal
timing was important for understanding depression and anxiety (Buchanan et al.,
1992) whereas another concluded that the findings in the literature were too spotty to
draw such a conclusion (Alsaker, 1995b, 1996). But evidence is mounting from wellcontrolled longitudinal studies that early maturing girls have more anxiety, distress,
and depressive feelings (Aro & Taipale, 1987; Dubas, 2002; Ge, Conger, et al., 1996,
2001; Graber et al., 1997; Stice, Presnell, & Bearman, 2001; Wichstrom, 1999; 2000;
Zehr, Culbert, Sisk, & Klump, 2007) and have a poorer body image (Muris, Meesters,
van de Blom, & Mayer, 2005; Williams & Currie, 2000). In one study, early maturing
girls experienced more depressive symptoms than on-time and late maturing girls
every year from the 7th to the 12th grade (Ge, Conger, et al., 2001), and at least two
studies have found associations between early puberty and suicidal behavior
(Wichstrom, 2000; Stattin & Magnusson, 1990). Others have shown that advanced
pubertal maturation at age 11-12 was linked to experiencing stress but not other types
of psychological difficulties (Simon, Wardle, Jarvis, Steggles, & Cartwright, 2003). In
a recent cross-sectional study, early pubertal timing predicted higher emotional
arousal which in turn predicted increased depressive affect (Graber, Brooks-Gunn, &
7
8Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
Warren, 2006). At the same time it should be said that when significant associations
between pubertal timing and depressive affect are reported, the effects are low to
moderate, and they differ markedly between studies (Ge, Conger, et al., 1996;
Wichstrom, 1999). Scholars have attributed this to power problems, noting that
significant effects are more likely to appear in studies with large rather than small
samples (Alsaker, 1996; Stice et al., 2001). The most reasonable conclusion seems to
be that early pubertal timing is associated with depressive mood and poor body image
in mid-adolescence, but perhaps not with a wider spectrum of internal problems (Stice
et al., 2001). If the timing hypotheses were correct, however, and the link between
early maturation and problem behavior were due to the distress caused by being out of
synch with peers, one would expect stronger, more robust links between early timing
and a variety of types of internal distress.
Part of the reasoning involving internal distress is the idea that the weight gain
associated with puberty, particularly for early developers, is one of the things that
produces distress. Body mass is greater on average in early than late maturing girls
(Haynie, 2003; Johansson & Ritzén, 2005), and early maturing girls tend to be more
concerned and dissatisfied with their bodies, particularly their weight (Caufman &
Steinberg, 1996; Crockett & Petersen, 1987; Dorn, Crockett, & Petersen, 1988; Ge,
Elder, Regnerus, & Cox, 2001; Graber et al., 1997; Stattin & Magnusson, 1990; Stice
et al., 2001). Some suggest that this might produce body shame and disordered eating
(Graber, et al, 1997; Keel, Fulkerson, & Leon, 1997; Lindberg, Grabe, & Hyde, 2007)
and, indeed, a link between early puberty and disordered eating has been found (Zehr
et al., 2007). Earlier reviews of the literature have considered the association between
poor body image and early pubertal timing to be robustly established (Alsaker, 1995a,
1996; Buchanan et al., 1992). But in view of the high stability of body mass from
childhood to late adolescence (McLaren, Hardy, & Kuh, 2003), the question is
whether the stronger body dissatisfaction among early-maturing girls in midadolescence holds after controlling for their body mass before puberty. A recent
prospective longitudinal study (Stice & Whitenton, 2002) and another using
retrospective information about prepubertal weight (Ackard & Peterson, 2001) failed
to confirm that early menarche was linked to later low body dissatisfaction. However,
another longitudinal study found a low but significant correlation between age at
menarche and satisfaction with weight at ages 26 and 54 after controlling for age 7
body mass (McLaren et al., 2003). Even when controlling for body mass at ages 7 and
54, age at menarche predicted poorer body satisfaction at age 54. To date, however,
the evidence from longitudinal studies is conflicting. Despite a wealth of studies
reporting associations between body image and pubertal timing (see Alsaker, 1996),
only a few studies have controlled for prepubertal body mass, and they have yielded
conflicting results. Because of this, the issue cannot be considered settled. Thus, it
appears that although there are provocative findings from some recent studies, internal
distress has not yet been conclusively linked with either early timing itself or the
weight gain that accompanies it.
The remaining question is whether the distress that has been reported to be
associated with early pubertal timing in some studies leads to normbreaking behavior.
There are several studies that have shown relations that would be expected if this is
true, although they have not directly tested the idea. One study suggested that distress
is associated with entering mixed sex peer groups and receiving negative reactions
from parents, but that for early maturing girls distress can also be a predictor of future
engagement with deviant friends (Ge, Conger, et al., 1996), which supports the idea
that distress is involved in the link between early puberty and problem behavior.
8
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
9
PAPER I
Furthermore, early maturing girls have been found to have three times higher risk for
having comorbid depression and substance use compared to other girls (Stice et al.,
2001), which also supports the idea. These findings suggest that daily living
conditions in their peer groups contribute to the early maturing girls’ depressed mood.
In a well-designed longitudinal study, gender differences in depression throughout
adolescence were reduced to nonsignificance once pubertal timing, stressful life
events, and the interaction between pubertal timing and life events, were controlled
(Ge, Conger, & Elder, 2001). Early maturing girls were more likely than late maturing
girls to encounter stressful life experiences, and early pubertal timing in interaction
with stressful life events predicted depressive symptoms in middle adolescence, but
this does not necessarily explain why early maturing girls have more problem
behavior. On the other hand, two studies have directly tested the idea that emotional
distress or depressive symptoms explain the link between early puberty and problem
behavior (Weisner & Ittel, 2002; Tschann et al., 1994). Although both studies showed
that distress was linked to substance use, they failed to show that distress was the
mechanism linking early puberty and problem behavior. Overall, these studies reveal
that whether pubertal timing is associated with adolescent depression and distress
depends on the girls’ everyday life experiences. But the evidence for a causal link
from internal distress to normbreaking behavior seems to be weak.
The off-time and early timing hypotheses are built on the idea that emotional
distress influences psychosocial adjustment. This is a directional hypothesis.
However, it is far from firmly established. There is still little in the empirical literature
that supports the claim that internal states like stress, low self-esteem, and depression
of the early maturing girls generally drive problem behavior. One plausible,
alternative hypothesis is that it works the other way around. Internalizing problems
such as depressive affect might be a consequence of problem behavior for early
maturing girls. For example, as a consequence of engaging with older peer networks,
early maturing girls may become marginalized in their same-age peer groups. As a
consequence of deviant peer association and problem behavior, they may encounter
more negative reactions from parents at home and peers at school. Poor peer
relationships are strongly associated with depressive mood (Patterson & Stoolmiller,
1991), so it is reasonable to expect that early maturing girls might develop emotional
problems and depressive mood if they feel different from their age mates and get
negative reactions from them. Another interpretation is that the issue of body image
becomes salient when early developed girls become interested in boys (Halpern,
Udry, Campbell, & Suchindran, 1999). With puberty comes an increase in body fat,
and early developed girls typically report poorer body images than later developed
girls (Stattin & Magnusson, 1990). Early pubertal development has been associated
with romantic involvement, as well as with poor body image and depression among
sixth grade girls (Compian, Gowen, & Hayward, 2004), and romantic involvement
has also been associated with poor body image and distress, suggesting that romantic
involvement is linked to negative outcomes for girls. Overall, the explanation from
the early timing hypothesis that internal distress of being out of sync with peers leads
to problem behavior is still not well founded. Future longitudinal studies are needed to
contrast these two hypotheses about internal states as a cause or as a consequence of
differences in psychosocial adjustment between early and late maturing girls.
9
10
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
Peer Influence. Another explanation for the link between early maturation
and problem behavior, the peer socialization hypothesis, places no emphasis on
internal distress, but focuses instead on the implications that early puberty has for peer
relationships (Magnusson, Stattin, & Allen, 1985, 1986; Stattin & Magnusson, 1990;
Caspi et al., 1993). According to this hypothesis, early maturing girls will tend to
affiliate with peers who are similar to them in biological maturation, and these will be
chronologically older peers. In addition, they will begin dating earlier than their peers,
and their boyfriends will also be similar to them in biological maturity, so they will be
chronologically older. Because older adolescents are on average engaged in more
normbreaking than younger adolescents (Sampson & Laub, 2003), girls who join
older peer groups will have peers who are doing more normbreaking than their sameage peers would do, and these older peers will socialize them into the same kinds of
behaviors. Thus, according to this hypothesis, advanced peers make early-maturing
girls vulnerable to developing problem behavior. This should be especially true in
high-risk neigborhoods with an abundance of deviant adolescents.
This explanation rests on established findings about peer relationships in
adolescence. Peers become increasingly important in adolescence, and adolescents
spend progressively more time with friends and romantic partners as they move
through adolescence (see Zimmer-Gembeck, 2002, for a review). Adolescents select
friends who are similar to them, a phenomenon known as homophily (Dishion,
Patterson, & Griesler, 1994; Hartup, 1996; Kandel, 1978; 1986), and then they
become more similar to their friends over time, which seems to be a peer influence
effect (Kandel, 1978; 1985). These same phenomena should apply to early maturing
girls. If so, then girls who mature early will seek out friends whom they perceive as
similar to them in maturity, and those will be older girls and other early developers.
They will also establish romantic relationships with boys earlier than their same-age
peers, most often with older boys, because they perceive older boys as similar to them
in maturity (Rowe & Rodgers, 1994; Zelnik & Shah, 1983). Thus, the normal process
of developing romantic relationships in adolescence will come earlier for early
developing girls, and the normal process of being attracted to similar others will lead
early maturing girls into relationships with peers and romantic partners who are older
than they are.
According to the peer socialization hypothesis, these changes of peer groups will
influence the behavior of the early maturing girls for the following reasons. Boys are,
on average, more involved in normbreaking than girls, and older adolescents, boys
and girls alike, are more involved in normbreaking than younger adolescents. Thus, if
a girl enters into a boyfriend’s group of friends, or generally engages with older peers,
her peer group is likely to be more normbreaking in terms of drinking, drug use, and
delinquency than the peer groups of girls who associate mainly with those of the same
age and gender. The girl with older peers or an older boyfriend or both will more
often find herself in leisure-time settings in which drinking, drug use, and delinquency
are normative. Indeed, as shown in the original work on the peer socialization
hypothesis (Stattin & Magnusson, 1989, 1990) and in later studies (Costello, Sung,
Worthman, & Angold, 2007; Dick et al., 2000; Ge, Conger, et al. 1996; Haynie, 2003;
Lynne et al., 2007; Mezzich, Giancola, Lu, Parks, Ratica, & Dunn, 1999; Patton et al.,
2004; Silbereisen, Petersen, Albrecht, & Kracke, 1989; Silbereisen & Kracke, 1993;
Wichstrom, 2001) early maturing girls associate more with deviant or normbreaking
peers, including adult male sex partners, in mid-adolescence than do on-time or late
maturing girls. Moreover, early maturing youth are more likely than later maturers to
have obtained alcohol from other adolescents (Storvoll, Pape, & Rossow, 2008).
10
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
11
PAPER I
According to the peer socialization hypothesis, this is largely because the peers of
early maturing girls are older than those of on-time and later maturing girls. Thus,
according to this explanation, the link between early development and normbreaking
appears because early-maturing girls seek out similar peers in the same way that
everyone else does, but in so doing, they find themselves in older, more socially
advanced peer groups and they adopt that behavior.
Although the peer socialization hypothesis takes into account relationships with
peers of both genders, romantic relationships with boys play a more important role in
the explanation. These are potentially the most negative influences because on
average boys are higher than girls on normbreaking at every age. Consequently, a
relationship with an older boy is a peer relationship in which the peer’s level of
normbreaking is theoretically the most discrepant from the girl’s, and as such, it offers
the most potential for negative socialization. Thus, an important part of the peer
socialization explanation is the assumption it makes about romantic relationships with
boys.
There are two premises to the peer socialization hypothesis: (a) that pubertal
timing is associated with an early onset of romantic and sexual relationships, and (b)
that girls with early romantic and sexual relationships are more likely to engage in
antisocial behavior by virtue of entering an older boy’s more normbreaking peer
group. Also, if there is a mediating role of social relationships, then limited
association should exist between age at menarche and problem behavior once
individual differences in romantic and sexual relationships are taken into account. The
question is whether the literature supports these.
Concerning the premise that early puberty means early establishment of romantic
relationships with boys, many studies have reported that by mid-adolescence early
maturing girls have more experience than their same-age peers with different aspects
of romantic relationships with boys and sexual development (feelings of being in love,
dating, sexuality, going steady, viewing sexual media, abortions, and being subject to
sexual abuse). This appears in studies done almost six decades ago (Stone & Barker,
1937, 1939) and those during the past two or three decades in North American and
European samples (Aro & Taipale, 1987; Baumeister, Flores, Marín, 1995; BrooksGunn, 1987; Brown, Halpern, & L’Engle, 2005; Cavanagh, 2004; Crockett &
Petersen, 1987; Flannery, Rowe, & Gulley, 1993; Garwood & Allen, 1979; McCabe,
1984; Meyer-Bahlburg, Erhardt, Bell, Cohen, Healy, Feldman, et al., 1985; Friedman,
1992; Goodson, Evans, & Edmundson, 1997; Haynie, 2003; Jorm, Christensen,
Rodgers, Jacomb, & Easteal, 2004; Phinney, Jensen, Olsen, & Cundick, 1990;
Richards & Larson, 1993; Rodriguez-Tomé, Bariaud, Cohen Zardi, Delmas,
Jeanvoine, & Slylagyi, 1993; Rowe, Rodgers, & Mesek-Bushey, 1989; Schor, 1993;
Silbereisen & Kracke, 1993, 1997; Simmons & Blyth, 1987; Simmons, Blyth, &
McKinney, 1983; Simmons et al., 1979; Smith, Udry, & Morris, 1985; Turner, Runtz,
& Galambos, 1999; Stattin & Magnusson, 1990; Udry, 1979, 1988; Udry & Billy,
1987; Zelnik, Kantner, & Ford, 1981). It also appears in nonwestern cultures such as
Hong Kong (Lam, Shi, Ho, Stewart, & Fan, 2002) and North Sudan (Otor & Pandey,
1999). The link between pubertal status and romantic relationships has been
established as early as at 10-13 years (Compian et al., 2004). Furthermore, teenagers
themselves tend to link early maturing with opposite-sex relationships (Faust, 1983).
With puberty there is an increasing tendency for girls, but not for boys, to view one's
body as an object for others to look at and evaluate (Lindberg et al., 2007) and when
girls think about physical maturation, they are aware that their physical development
will arouse interest from boys and that there are sexual issues connected with their
11
12
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
physical development (O’Sullivan, Heino, Meyer-Bahlburg, & Watkins, 2000). Thus,
the link between pubertal development and initiation of sexual relationships seems to
be quite robust, and it seems well established that girls who mature early enter into
romantic and sexual relationships earlier than their same-age peers.
Concerning the premise that developing relationships with boys will lead to higher
normbreaking for girls, there is ample evidence in the adolescent literature for a link
between early romantic interests and behavioral problems, even though the
mechanism linking them remains unclear. Girls’ early sexual transition and early
engagement with boys have well-established associations with health problems and
behaviors such as delinquency, drug and alcohol use, smoking, having deviant peers,
low school motivation, dropping out of school, truancy, disobeying parental rules,
running away from home, pregnancy, and parental conflict (Bingham & Crockettt,
1996; Chilman, 1986; Costa, Jessor, Donovan & Fortenberry, 1995; Elliot & Morse,
1989; Fidler, West, Jarvis, & Wardle, 2006; Gibbs, 1986; Giordano, 1978; Gustafson
& Magnusson, 1991; Gustafson, Stattin & Magnusson, 1992; Haynie, 2003; Jessor &
Jessor, 1975; Jessor, 1992; Ketterlinus, Lamb, Nitz & Elster, 1992, Ketterlinus &
Lamb, 1994; Metzler, Noell, & Biglan, 1992; Rosenbaum & Kandel, 1990; Zabin,
1984). As these problems often cluster together rather than appearing isolated, they
have been discussed as lifestyles, and have been empirically investigated as
constellations or patterns of problems (Jessor & Jessor, 1977; Magnusson & Stattin,
1998; Stattin & Magnusson, 1996). In various theoretical models, adolescent lifestyles
linked to health endangering behavior often involve early and unsafe sexuality in
combination with normbreaking, school problems, smoking, and drug and alcohol use
(Jessor, 1992; Petersen, Leffert, & Graham, 1995). Thus, although early sexual
involvement is linked to problem behavior, it is not clear whether it is one more
example of a problem behavior or, as the peer socialization hypothesis claims, a
channel through which girls become exposed to peers who socialize them into other
problem behaviors. This is a question about pre-existing characteristics or third
variables.
In terms of direct tests of the peer socialization idea, several longitudinal studies
have shown empirical evidence for this explanation (Caspi, 1995; Ge, Jin et al., 2006;
Halpern, Kaestle, & Hallfors, 2007; Haynie, 2003; Lynne et al., 2007; Stattin &
Magnusson, 1989, 1990). Early maturing girls have more socially advanced peer
networks, including older peers and boyfriends and they also have peers who are more
deviant or normbreaking than the peers of other girls (Haynie, 2003), which is one
premise of the peer socialization hypothesis. Findings about peers have concerned
both deviant peers and romantic relationships. There are also studies that have directly
tested that peer socialization idea. In one study, the early maturing girls’ problem
behavior was largely due to the greater romantic interests and relationships (Stattin &
Magnusson, 1989, 1990). In another study, controlling for relevant selection factors,
higher levels of self-reported delinquency and delinquent peers were found for earlymaturing girls who had entered mixed-sex schools rather than all-girls schools (Caspi,
1995). In a third study, having delinquent peers at 6th grade, mediated the association
between perceived pubertal timing and delinquency and aggression in grades 6
through 8 (Lynne et al., 2007). Similarly, time spent with peers, involvement in
romantic relationships, and having deviant peers together with parental relations
mediated the link between pubertal timing and party deviance, which was measure of
smoking, drinking, lying to parents, school truancy etc. in another study (Haynie,
2003). In an additional study, having a romantic partner, particularly one that is older,
mediated the link between perceived physical maturity and high risk behavior
12
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
13
PAPER I
(substance use and risky sexual behavior; Halpern et al., 2007). Finally, among early
adolescent African Americans, early-maturing girls who had substance-using peers,
were more willing to and had a more intentions to use substances themselves two
years later than did other girls with substance-using peers (Ge, Jin, et al., 2006). All of
these studies found that peers, including older peers, boys, and romantic partners,
mediate the link between pubertal timing and problem behavior. Thus, although the
longitudinal studies that have examined the role of peers in the link between early
puberty and problem behavior have had different foci, all findings have been in line
with the peer socialization hypothesis.
One limitation of most of these studies is the use of girls’ reports of their peers’
behavior rather than independent reports from the peers themselves. Because the
knowledge of peers’ behavior in these studies has all come through the eyes of the girl
herself, they are subject to the false consensus effect, in which people assume that
others are more like them than they actually are (Ross, Greene, & House, 1977).
Studies suggest that self-reports of peers’ behaviors are more strongly associated with
individuals’ own behavior than are independent measures of peers’ behaviors (c.f.
Iannotti, Bush & Weinfurt, 1996). Adolescents might project their own behavior onto
friends or justify or rationalize their own behaviors by misrepresenting their friends’
deviance (Conger & Rueter, 1996; Kandel, 1986; Urberg, Degirmencioglu, & Pilgrim,
1997). Hence, data from self-reports of peer behavior are likely to systematically
overestimate the role of peer association in individual development (Kandel, 1985).
The consequence for understanding the link between early maturation and problem
behavior is that even with longitudinal data using data for girls’ perceptions of their
friends and romantic partners, one cannot know whether peers socialize girls into
normbreaking or whether normbreaking girls see their peers as similar to themselves
and thus higher in normbreaking than they actually are. Because peers’ behavior is a
critical link in the explanation, research on this explanation should incorporate
independent measures of peers’ behavior.
In sum, the peer socialization hypothesis offers an advantage of being based on
established principles of peer affiliation and influence. There are large literatures that
seem to support the basic premises, although they are open to alternative
interpretations. The studies that have examined the hypothesis directly, as well as
other studies, have been supportive. There are some basic issues that are still in
question, however. One of the most critical is whether romantic relationships with
boys play a mediating role or are themselves problem behaviors that can be attributed
to some other cause. Another is whether independent reports of peers’ behaviors will
yield the same results as past studies in which this information has come from the
girls themselves. A third is whether the links between pubertal timing and problem
behavior could be more easily explained by conditions already present at the time of
puberty. These issues must be addressed before the processes laid out in the peer
socialization hypothesis can be considered established.
Context effects. The most recent attempts to explain the link between
early pubertal timing and problem behavior, both externalizing and internalizing, have
focused on contextual conditions (Ge, Natsuaki, Jin, & Biehl, 2007). According to the
contextual amplification hypothesis, contextual conditions moderate the effects of
pubertal maturation on girls’ problem behavior. Of interest are the contexts that
youths are directly and frequently in contact with, such as family, peers, and
neighborhood. According to the contextual amplification hypothesis, these
psychosocial contexts could either be adverse or protective. It is assumed that early
pubertal maturation is particularly detrimental for girls in adverse psychosocial
13
14
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
contexts, who, for instance, live in disadvantaged neighborhoods or experience harsh
parenting. For early maturing girls in supportive environments, however, the chances
of problems are low. The contextual amplification hypothesis views the wellestablished link between early puberty and problem behavior as either a result of
cumulative risks, where early puberty is one and an adverse context is another, or it is
described using diathesis-stress models where early puberty is a diathesis and an
adverse context is the stress. Thus, according to this hypothesis, contextual factors do
not themselves explain the link between pubertal timing and problem behavior, but
they moderate the effects described in the other hypotheses. As such, the contextual
amplification hypothesis is compatible with all the previously described hypotheses.
There is mounting empirical support for the contextual amplification hypothesis.
For instance, whether early developed girls will be more involved in problem
behaviors has been found to be dependent on type of school (Caspi et al., 1991; Caspi,
1995), family (Ge, Conger, & Elder, 1996), neighborhood (Ge et al., 2002), and the
broader community (Dick et al., 2000). These findings suggest that contextual
circumstances increase or decrease the risks that are associated with early puberty.
Although the mechanisms through which early puberty is linked to problem behavior
are not specified in the contextual amplification hypothesis, deviant peers have
frequently been mentioned as an important feature of a context that increases the
negative impact of early puberty in this literature (Caspi et al., 1993; Ge et al., 2002;
Obeidallah et al., 2004). In this way, the contextual amplification hypothesis is
consistent with the peer socialization hypothesis.
Summary and Purpose of This Work
Although the link between early pubertal timing and problem behavior is well
established, the mechanisms to explain this link are less well understood. The
hypotheses based on the stress of being off time are mainly relevant for explaining a
link to internalizing problems, and this link is less robust than for externalizing-type
problem behavior. In addition, the mechanisms through which internal problems
might be transformed into external problems are not clearly spelled out in these
hypotheses. The accentuation hypothesis has received support in one longitudinal
study and has the advantage of proposing a general mechanism for the accentuation of
problem behavior, but why the accentuation of problems should be more true for early
developing girls than later developing girls has not been spelled out. The contextual
amplification hypothesis does not by itself explain the mechanisms linking early
puberty with problem behavior, but it has the advantage of being compatible with all
the other hypotheses and offering a way to understand why results might be
inconsistent across samples. The peer socialization hypothesis has the advantage of
being based on established peer selection and socialization principles. But although it
is widely accepted and used as an aid to the interpretation of findings, few studies
have been designed to test it directly by specifying the mechanisms involved and
making testable predictions about them. Our goal in this work has been to specify the
theoretical mechanisms involved in the peer socialization explanation of the link
between girls’ early maturation and problem behavior and to test this explanation
using ideas about contextual amplification to predict which contexts should facilitate
and which should inhibit peer socialization. In this way, we unite the major ideas of
the contextual amplification and the peer socialization hypotheses.
For the present purposes, we propose an extension of the contextual amplification
hypothesis. The work thus far has focused on adverse contexts, and shown
amplification of the link between pubertal timing and problem behaviors in adverse
14
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
15
PAPER I
contexts. We propose that contexts might not be considered adverse, but they might
amplify the link between pubertal timing and problem behavior by making it easy for
early developing girls to socialize with older peers, particularly boys, who might draw
them into problem behavior. For instance, cultures with liberal attitudes toward
adolescent romantic and sexual contacts are not necessarily adverse, but one could
argue that they provide conditions in which peer socialization of early-developing
girls should be amplified. This is an extension of the contextual amplification
hypothesis but it is commensurate with the basic contextual amplification idea. We
propose, then, that negative peer socialization of early developing girls should be
amplified in contexts that bring together youths of different ages and diminished in
contexts that keep youths of different ages separated. It should be amplified in leisure
contexts where a lot of troubled youths spend their time and diminished in contexts
where most youths are well adjusted.
In this monograph, we report a programmatic series of studies designed to test the
peer socialization hypothesis within different contexts (cultural, school, and
neighborhood). In Chapter II, we test cultural level predictions. We use samples from
two different cultures—Sweden, in which the attitudes toward teenage romance and
sexuality are tolerant, and Slovakia, in which the attitudes are quite restrictive. If
contact with boys mediates the link between early puberty and problem behavior, then
in Slovakia, where heterosexual contacts are restricted, there should be no link
between pubertal timing and problem behavior. In Sweden, where heterosexual
contacts are allowed, pubertal timing should be linked to problem behavior. In
Chapter III, we turn to more proximal, everyday contexts—in-school versus out-ofschool contexts. We examine the basic assumptions of the peer socialization
hypothesis by making predictions about these two contexts. According to the peer
socialization hypothesis, early maturing girls select friends and romantic partners who
are similar to them, namely older peers. Our argument is that at school they are
restricted from affiliating with older peers because of the age-graded nature of school
activities; in contexts outside of school, they are freer to associate with older peers.
The prediction from the peer socialization hypothesis is that early- and laterdeveloping girls should have similar in-school friends, but the out-of-school friends of
early-developing girls should be older and higher on delinquency than the out-ofschool friends of later-developing girls. In Chapter IV, we focus on a leisure
context—the Swedish neighborhood youth centers—where delinquent youths are
known to congregate (e.g., Mahoney, Stattin, & Magnusson, 2001). According to the
peer socialization and contextual amplification hypotheses, if early-developing girls
spend a lot of time in contexts where delinquent boys hang out, they should be more
likely to be socialized into delinquency than early-maturing girls who do not spend
time in those contexts.
In this study, we predict that girls who (a) develop early, (b) attend the youth
centers, and (c) are romantically involved with boys will have the highest levels of
self-reported delinquency in adolescence and registered criminality in adulthood. In
Chapter V, we ask whether differences in problem behavior between early and later
developing girls can be explained by differences that already existed before puberty.
If they can, then early pubertal timing does not usher in new behaviors, but transfers
the differences that existed in childhood to adolescent life. We test this idea by
examining individual differences in problem behavior at age 10, when almost all girls
were pre-menarcheal. Chapter VI provides a discussion. In all, there are three
longitudinal and two cross-sectional cohorts used in these studies from the 1970s,
1990s, and 2000s. We predict and find peer socialization differences between contexts
15
16
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
on different levels: between two specifically chosen cultures, between school and
leisure settings, and between different leisure contexts. In so doing, we provide
evidence for both peer socialization and contextual amplification.
16
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
17
In this chapter, we use samples from two cultures, which were chosen as being
cultural contexts that, theoretically, should amplify or diminish the peer socialization
effect. Previous studies in the pubertal timing literature have considered culture in
reference to whether pubertal development has similar implications across ethnic
groups and social classes (cf. Clausen, 1975; Siegel, Yancey, Aneshensel, & Schuler,
1999) and whether the findings that have appeared in Western countries also apply to
nonwestern countries (Lam et al., 2002). They have also examined differences
between countries and over time in pubertal development (cf. Tanner, 1989). They
have not, however, used theoretically chosen cultures to test a specific hypothesis
about mechanisms in the link between pubertal timing and problem behavior. We
reason that if the peer socialization hypothesis is correct, then early maturing girls
should develop problem behaviors only if they are able to associate with peers who
have higher levels of problem behavior than they. Older girls might draw early
maturing girls into problem behavior, but he most important peers should be boys,
because boys have higher levels of problem behaviors than girls at every age. If, for
one reason or another, early maturing girls are prevented from more problematic
peers—particularly boys—their behavior should not differ from later maturing girls.
In this study, we test this idea directly by selecting samples from societies that differ
in the extent to which adolescents have freedom to form romantic relationships with
boys. Thus, we test the peer socialization hypothesis on the cultural level.
We compare two countries—Sweden and Slovakia—which, according to what we
know, differ considerably on people’s views of adolescent sexuality. Slovakia is a
socially conservative, Catholic society where socialization in general tends to be more
collectivistic than individualistic and views about adolescent sexuality are quite
restrictive. Among young adolescents, heterosexual contacts are strictly limited, and
sexual behavior is taboo. In contrast, Sweden is much more liberal about adolescent
sexuality and heterosexual contacts. Sexuality is considered part of the normal
transition to adult life. Sex education enters early in schools, contraceptives are
readily available to youths at government supported free health clinics, abortions have
been free and available by law since 1975, and adults are generally accepting of
opposite-sex contacts and sexual behavior among youths (Edgardh, Lewin, & Nilsson,
1999; Forrest, 1990; Stattin & Magnusson, 1990). There have never been formal
arrangements for adolescent heterosexual contacts, such as a classical dating system
or a generally accepted age at which youths should begin dating. Consequently,
people’s ideas about how old adolescents should be when they start having romantic
relationships are more individually determined than culturally agreed upon. Thus, in
Slovakia and Sweden there seem to be clear differences in the freedom that adolescent
girls have to form relationships with boys.
Because of these differences, we predicted that sexual activity should be more
prevalent among Swedish adolescents than among Slovakian adolescents, and if
romantic relationships with boys mediate the link between pubertal timing and
problem behavior, that link should be stronger in Sweden than in Slovakia.
Specifically, we made three predictions. First, because Slovakian society generally
discourages adolescent romantic relationships with boys, early pubertal development
should be associated with early romantic relationships with boys for Swedish girls but
not for Slovakian girls. We assumed, however, that some Slovakian girls would
succeed in developing relationships with boys despite the societal constraints. Thus,
we predicted, second, that in accordance with the peer socialization hypothesis, girls
PAPER I
Chapter II: The broader, cultural context
17
18
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
in both societies who had relationships with boys early in adolescence would be more
normbreaking than those who did not. Finally, we predicted that among Swedish girls,
early pubertal development would be linked to normbreaking behavior, but among
Slovakian girls it would not, because the mechanism linking the two—relationships
with boys in early adolescence—would be largely absent. In sum, we expected to find
a link between pubertal timing and norm-breaking in the Swedish sample, and we
expected it to be mediated by romantic relationships with boys. We expected no link
between pubertal timing and normbreaking in the Slovakian sample, because the
society does not tolerate heterosexual contacts, which would be initiated by puberty in
a less restrictive society such as Sweden.
We made one final prediction. We predicted that the findings obtained for the
Swedish girls would hold even for a Swedish sample of more than 30 years ago.
Tolerant attitudes toward adolescent sexual and romantic contacts have been present
in Sweden for several decades. For example, a couple of decades ago, Swedish
children had better knowledge of sexual issues than same-age children from other
countries (Goldman & Goldman, 1982). Thus, we assume that romantic relationships
with boys mediate the link between pubertal timing and problem behavior and that
this would hold today and much earlier.
Method
Participants. The participants were from two Swedish and one Slovakian
cohort. They all participated in broad surveys of self-experienced symptoms and
adjustment problems among 15-year-old girls. The first survey was done in a city in
central Sweden (Örebro) during the 1970-71 school years, and a repeat survey was
done in the same Swedish community and in a similar sized community in central
Slovakia community (Banská Bystrica) during the 1995-96 school year. For
simplicity, we shall refer to the two Swedish cohorts as “the Sweden 1970 cohort” and
“the Sweden 1995 cohort.”
The Swedish community had 100,000 inhabitants in 1970 and 120,000 inhabitants
in 1995. The 1970 cohort included all adolescent girls attending grade 8 in fall 1970.
These girls were enrolled in the longitudinal research program “Individual
Development and Adjustment” (IDA; Magnusson, 1988; for details of the 1970-71
data collections, see Magnusson, Dunér, & Zetterblom, 1975). Data collections were
made on two occasions: in fall (November, 1970) and in spring term (May, 1971).
The 1970 cohort included 590 girls, and 522 (88.5%) of them were present at school
and filled out a questionnaire in the fall term.
The 1995-96 survey in Sweden was done as an identical replication of the original
study. The Swedish 1995 cohort included all the adolescent girls attending grade 8 in
the same city in the 1995-96 school year. Data collections were made on two
occasions—November 1995 and May 1996—just as in the earlier study. The 1995
cohort included 602 girls, and 529 (87.9%) of them were present and filled out the
questionnaire during the November data collection.
Data were collected in 1995 (November) for the Slovakian participants. The
Slovakian cohort consisted of freshmen in all high schools in a town of about 90,000
in central Slovakia. There were 1018 girls registered in the town for the spring term.
Among these, data were collected for 955. Hence, the participation rate was 94% for
the Slovakian cohort. The average age of the girls was 14 years and 5 months in all
three cohorts at the November data collection.
Trained test leaders were used, and the pupils answered the questions in their
regular classrooms. The administration was conducted in a similar way for all cohorts
18
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
19
PAPER I
– the same instruments were used, and the instructions to the participants were
identical – and were made the same month (November) in both cohorts.
There is one difference that should be noted. The Slovakian girls filled out their
self-reported pubertal timing in November (the measure was included in the
Symptoms Questionnaire; see below); the Swedish girls reported on their pubertal
timing in May. Hence, the 1970 Swedish girls were 5 months older when they
reported on their age at menarche. We do not believe that this affects the results
reported.
Measures. Normbreaking measures. Measures of normbreaking behavior
were included in two questionnaires: the Symptoms Questionnaire (Magnusson et al.,
1975) administered in November 1970 and 1995, and in a Norm Inventory,
administered in May 1971 and 1996 for the Swedish girls (Magnusson et al., 1975).
The Slovakian girls reported on both these instruments in November 1995. In order to
have measures of normbreaking behavior taken at the same month of the year, we
used the measure of normbreaking behavior from the Norm Inventory when we report
comparisons between Swedish and Slovakian girls, and from the Symptoms
questionnaire, when we compare the 1970 and the 1995 Swedish cohort.
Normbreaking behavior used for the Swedish 1971-96 comparison was a scale
aggregated from eleven questions in the Symptoms Questionnaire about shop lifting,
alcohol abuse, signature forgery, pilfering, using hashish, vandalism, evasion of
payment, physically hitting somebody, harassment, and loitering in town in evenings
(Magnusson et al., 1975). The response scale was a 5-point scale: (1) no, (2) once, (3)
2-3 times, (4) 4-10 times, and (5) more than 10 times. The alpha reliability for the
scale was .75 in 1971 and .84 in 1996.Normbreaking behavior used for the crosscultural comparison was taken from the Norm Inventory (Magnusson et al., 1975).
The instrument contained eight norm violating activities at home (ignoring parents’
prohibitions, staying out late without permission), at school (cheating on an exam,
being truant), and during leisure time (smoking hashish, getting drunk, shoplifting,
loitering in towns in evenings). For each of these, participants reported on their own
evaluations The answers were given on 7-point scales with the alternatives: “I think it
is (1) very foolish, (2) foolish, (3) rather foolish, (4) not really OK, (5) rather OK, (6)
OK, and (7) quite OK.
The measure of the girls’ own norm violation consisted of answers to the question
how many times the girls had done each of the eight norm violating behaviors.
Answers were given on 5 point Likert scales with the alternatives: never (1), once (2),
2-3 times (3), 4-10 times (4), and more than 10 times (5) (see Magnusson et al., 1975,
pp. 100-103).
Behavioral intentions in connection with breaking norms. For each of these eight
measures in the Norm Inventory, descriptions of concrete instances were constructed.
Situations depicting the different types of norm violating activities were presented and
the girls were asked about their own behavioral intentions to breach respective norm,
on a scale with the alternatives (“What do you think you would do in this situation”):
would definitely not cheat (etc.) (1), would probably not (2), would perhaps not (3),
uncertain what to do (4), would perhaps (5), would probably (6), and would definitely
(7).
The eight norm violating activities for each of the different norm dimensions
(evaluations, behavioral intentions, and own norm violating behavior) were
aggregated into broader scales. The alpha reliability for the normbreaking behavior
(own behavior) scale was .84 for Swedish girls and .81 for the Slovakian girls. For the
two other scales (evaluations, behavioral intentions), the alphas varied between .84
19
20
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
and .90 for Swedish and Slovakian samples. The strongest difference in alpha
reliability was .05 units. Hence, very small differences in homogeneity for the
respective scales were found between girls in the two countries.
Other measures. For the other measures, we used the Symptom
Questionnaire, a 123-item instrument, which was originally constructed by the
investigators in the IDA research program (Crafoord, 1986; Magnusson et al., 1975).
The questionnaire covers different adjustment domains, such as physical symptoms;
emotional disturbances; self-esteem problems, behavior problems; problems in
relationships with parents, teachers, and peers; and sexuality.
Romantic relationships with boys. Three items measured how advanced the girls’
relationships to boys were. One was about perceived sexual maturation: “Do you feel
sexually more experienced than your age-mates? The response options were: never
(1), seldom (2), sometimes (3), rather often (4), and very often (5). Another question
was about intercourse: “Have you had sexual intercourse?” The response options
were: no (1), yes, once (2), and yes, several times (3). The third question was about
having a steady boyfriend: “Have you now or have you had a steady relationship to a
boy?” The response options were: have never had, don’t want to have (1), have never
had, but want now (2), have had earlier, but not now (3), have now, but not earlier (4),
and have now, and have had earlier (5). The last item was recorded into a
trichotomized measure in the calculations. Scale values 1 and 2 were coded 1, value 3
was coded 2, and values 4 and 5 were coded 3. The alpha reliability for this 3-item
measure in 1971 was .70 and it was .74 for the 1996 sample.
Pubertal timing. Puberty can be measured by various indicators, such as peak
height velocity, menarche, breast development, weight increase, body hair, or other
measures of secondary sexual characteristics. In the literature, age at menarche is one
of the most commonly used measures for establishing individual differences in
pubertal development in girls (see Dorn, Dahl, Woodward, & Biro, 2006 for a review
on issues of measuring puberty). It has been considered the most salient single
indicator of pubertal development for girls, but is certainly not an inclusive measure
of pubertal maturation in females (Dorn et al., 2006). Throughout the studies
presented here, we will use self-reported age at menarche as an indicator of pubertal
timing. Apart from its salience for girls, concerning reliability, self-reported
menarcheal age has consistently in the literature shown high test-retest stabilities
(Kaprio, Rimpelä, Winter, Viken, Rimelä, & Rose, 1995; Gilger, Geary, & Eisele,
1991), and the association between recalled menarcheal age and actual menarcheal
age that has been reported in many previous studies are substantial (over .60 for adult
samples, Bergsten-Brucefors, 1976; Gilger et al., 1991). As to validity, comparing this
measure with data for skeletal age and with height and weight measures, Stattin and
Magnusson (1990) reported satisfactory validity for the scale. We are aware that we
do not capture the complexity of pubertal development by relying on one single
measure, but our arguments is that age at menarche captures an essential aspect of
physical pubertal development among girls; refers to a discrete event rather than a
process over time (Pickles, Pickering, Simonoff, Silberg, Meyer, & Maes, 1998);
should be an event that girls remember easily (Coleman & Coleman, 2002); should be
less affected by the individuals’ subjective conceptions of their developmental stance
relative to age mates, as will be the case with a perceived pubertal timing measure
(Dick, Rose, Pulkkinen, & Kaprio, 2001); and can be used without involving parents
and school officials.
The age at menarche was asked in the Symptom Questionnaire. It was originally
coded in 6 categories: before age 10 (1), between 10 and 11 (2), between 11 and 12
20
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
21
Results
Stability Over Time in the Swedish Context. Our argument was that
because Swedish society has traditionally been tolerant of adolescent romantic
relationships with boys, early maturing girls have the opportunity to develop the kinds
of peer associations that eventually could lead them into normbreaking behavior. If
this is so, then the links between early maturation, romantic relationships with boys,
and normbreaking should be rather stable over time in Swedish society. To
demonstrate this, we look at these relations in two cohorts of Swedish 15-year-old
girls who were assessed 25 years apart in one city. As shown in the top row of Table
2.1, pubertal timing was associated with problem behavior in both the 1970 and the
1995 cohorts. The table also reveals that in both cohorts, the earlier the girls matured,
the more likely they were to perceive themselves as more sexually matured than
others their age, to be going steady, and to have had sexual intercourse. In both
cohorts, normbreaking behavior was also significantly associated with the three
measures of romantic relationships with boys. Apparently, the role of pubertal
maturation for romantic relationships with boys and problem behavior among girls in
this community is about the same today as it was a quarter of a century ago. Thus,
these initial findings are consistent with the assertion that the conditions in Swedish
society have, over many years, allowed early maturing girls a fair amount of freedom
to develop romantic and sexual relationships with boys.
Table 2.1.
Correlations relating pubertal timing and normbreaking behavior to different aspects of
romantic relationships with boys in 1971 and 1996 in one Swedish community
_________________________________________________________________________
1971
1996
Pubertal
Pubertal
timing
Normbreaking
timing
Normbreaking
_________________________________________________________________________
Normbreaking
-.23 ***
-.20 ***
(524)
(510)a
Sexually more matured
.27 ***
(498)
.39 ***
(507)
.25 ***
(519)
.36 ***
(523)
Steady dating
.26 ***
(498)
.29 ***
(507)
.22 ***
(517)
.34 ***
(521)
Intercourse
.28 ***
.46 ***
.28 ***
.41 ***
(502)
(511)
(522)
(525)
________________________________________________________________________
a
degrees of freedom; ***p < .001
PAPER I
(3), between 12 and 13 (4), after age 13 (5), and have not yet had my first period (6).
In all our studies presented here, we will use a condensed four-scale measure,
differentiating between girls who had their menarche: before 11 years (1), between 11
and 12 (2), between 12 and 13 (3), and 13 and later (4). The coding highlights
variations in early timing – in commensurate with our theoretical interest – rather than
covering the whole age variation of pubertal timing.
21
22
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
We ask, next, whether romantic relationships with boys seem to mediate the link between
pubertal timing and normbreaking to the same extent in both cohorts. To answer this
question, we first regressed pubertal timing on norm-breaking. If romantic relationships
with boys mediate the link between pubertal timing and normbreaking, then pubertal
timing should not be a significant predictor in this model once romantic relationships are
controlled. As reported in Table 2.2 (Model 1), pubertal timing was a significant
predictor of norm-breaking both in the 1971 and 1996 cohorts. In the 1971 cohort, when
sexual maturation, steady dating, and intercourse were entered into the equation, puberty
was no longer a significant predictor of norm-breaking. The beta weight was reduced
from -.24 to -.08. We tested whether this reduction was significant by determining
whether the slope (b) changed enough to move it out of its 95% confidence interval. The
95% confidence interval for b was -.081 to -.177; the new b of -.043 was outside of the
original confidence interval. We conclude that the association between romantic relationships
and normbreaking in the 1971 cohort was entirely mediated by romantic relationships.
In the 1996 cohort, pubertal timing was still a significant predictor (p < .05) after
sexual maturation, steady dating, and intercourse were entered into the equation, but the
beta weight was reduced from -.22 to -.09. The 95% confidence interval for b was -.096 to .214. With the three measures of romantic relationships in the equation, the b of -.064 was
outside the original confidence interval. We conclude that the effect of pubertal timing on
norm-breaking was significantly reduced in the 1996 sample once romantic relationships
were controlled. Thus, the mediational model was largely supported in both cohorts.
Taken together, these findings support the assertion that early maturing girls are more
involved in romantic relationships with boys and this largely explains their involvement
in problem behavior. Moreover, the findings show that this has been a stable
phenomenon since the 1970s. Thus, the tolerant attitudes toward adolescent sexuality in
Sweden seem to make it possible for early developing girls to form romantic
relationships with boys, and these relationships explain, statistically, the link between
early maturation and normbreaking.
22
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
23
Model 1.
Pubertal timing
Pubertal timing
R2
Model
Model 2.
Pubertal timing and romantic
relationships with boys
Pubertal timing
Sexual maturation
Steady boyfriend
Intercourse
R2
Model
+
p < .10 * p < .05
19711
19962
beta
beta
-.24 ***
-.22 ***
.06 ***
.05 ***
F(1,478) = 28.63*** F(1,509) = 26.63***
-.08 ns
.18 ***
.07 ns
.33 ***
-.09 *
.17 ***
.15 **
.22 ***
.26 ***
.23 ***
F(4,478) = 42.32*** F(4,509) = 36.89***
** p < .01
*** p < .001
Does the Societal Context Matter? The process described in the peer
socialization hypothesis depends on early developing girls being able to form
relationships with boys. If a society is intolerant of teenage romantic and sexual contacts,
early maturing girls should be prevented from forming relationships with boys. Such a
society should inhibit rather than amplify the peer socialization effect. Thus, if peer
socialization explains the link between early maturation and normbreaking, that link
should not appear in a society that is intolerant of teenage romantic or sexual contacts. To
test this prediction, we compare the Swedish and Slovakian cohorts from 1995.
Cultural similarity in age at menarche. First, to determine whether the age at
menarche was similar in the two cohorts, we compared the Swedish and the Slovakian
girls on the 4-point pubertal maturation scale. No significant differences were found (F
(1, 1467) = 0.28, p > .05). Most of the Swedish and the Slovakian girls had their first
menstruation between the ages of 12 and 13, and about one girl in ten had her first
menstruation before age 11. Specifically, 10.7% and 8.3% reported menarche before age
11 in the Swedish and Slovakian cohorts, respectively. Between ages 11 and 12, it was
25.5% and 24.3%; between ages 12 and 13, it was 39.2% and 46.4%; and at age 13 or
later, it was 24.6% and 21.1% in the Swedish and Slovakian cohorts, respectively. Thus,
the distributions of menarchal age were roughly similar in the two cohorts.
Cultural difference in prevalence of intercourse. A large literature suggests that
the timing of puberty largely determines the timing of intercourse. Nonetheless, we
expected that in the restrictive Slovakian society, 14-year-old girls would be less
likely to have had sexual intercourse than 14-year-old girls in the liberal Swedish
society. The results supported that assumption. At the average age of 14 years and 10
months, 14.6% of the Swedish girls reported having had intercourse (2.9% once and
11.8% on several occasions), whereas only 5.6 % of the Slovakian girls did (2.9%
once and 2.7% on several occasions) (2 (1) = 34.94, p < .001). Thus, the degree to
PAPER I
Table 2.2.
Predicting norm-breaking behavior from romantic relationships with boys and
pubertal timing in 1971 and 1996.
23
24
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
which pubertal timing is linked to romantic relationships with boys, which is the most
important condition necessary for the hypothesized peer socialization process to take
place, seems to be less in the Slovakian cohort than in the Swedish cohort. Next,
however, we examine this directly.
Differences in romantic relationships with boys between early and later
maturing girls. How did early and later maturing girls differ in heterosexual
experiences in the two societies? As reported in Table 2.3, the early maturing girls in
Sweden were considerably more advanced than their later maturing peers. The early
maturing girls in Slovakia also differed from their late maturing counterparts, but the
differences were less conspicuous. Two-way ANOVAS showed a significant country
by pubertal timing effect for perceiving oneself sexually more mature than peers (F
(3, 1423) = 7.65, p < .001) and for sexual intercourse (F (3, 1423) = 9.16, p < .001),
and a tendency for steady relationship (F (3, 1423) = 2.32, p = .07). In all three cases,
Swedish girls who were early matured were higher in these measures than later
developed Swedish girls and early and later developed Slovakian girls. We conclude
that the difference between early- and later-matured girls with regard to romantic
relationships with boys was considerably more pronounced for Swedish than for
Slovakian girls. Thus, the conditions are better in the Swedish sample than in the
Slovakian sample for early maturing girls to be socialized into problem behavior by
older boys.
Were the correlations between pubertal timing and different aspects of romantic
relationships with boys stronger for Swedish girls than for Slovakian girls? We tested
these differences with r-to-Z transformations. They are shown in the last column of
Table 2.3. The correlations between pubertal timing, on the one hand, and sexual
maturation, intercourse, and having a steady boyfriend, on the other were significantly
higher for Swedish girls than for Slovakian girls (sexual maturation: Z = 3.27, p <
.001; intercourse: Z = 4.01, p < .001; steady boyfriend: Z = 2.55, p < .05). We
conclude that the the correlations between pubertal timing and the different aspects of
romantic relationships with boys were stronger for Swedish females and for
Slovakian.
24
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
25
________________________________________________________________________
Age at menarche
Before
11-12
11
Swedish cohort
Sexual
maturation a
Intercourse b
Steady boyfriend c
Slovakian cohort
Sexual
maturation
Intercourse
Steady boyfriend
12-13
13+
Chi2
(3df)
Correlation with
pubertal
maturation
26.8
6.8
3.0
2.3
46.71 ***
-.24 ***
39.3
33.9
20.9
24.2
8.7
15.3
6.3
10.2
44.29 ***
18.91 ***
-.27 ***
-.19 ***
7.8
7.1
4.2
3.6
4.58 n.s.
-.07 *
12.8
20.5
5.8
23.9
3.2
19.3
6.6
16.6
13.09 **
3.74 n.s.
-.06 ns
-.05 ns
a
The percentage of girls who rather often or very often felt sexually more mature than
their same-age peers; b The percentage who had had sexual intercourse at 14:10 years; c
The percentage who had a steady boyfriend at 14:10 years; * p < .05; ** p < .01; *** p <
.001.
Cultural differences in the links between pubertal timing and normbreaking.
If associations with older, more normbreaking peers, particularly boys, explains why
early pubertal maturation tends to be linked to normbreaking, then pubertal timing
should be linked to normbreaking for Swedish girls more than for Slovakian girls,
because Swedish girls are more involved with boys. As reported in Table 2.4, this
expectation was confirmed. Pubertal timing was significantly associated with all three
normbreaking measures for girls in the Swedish cohort – evaluations, behavioral
intentions, and own normbreaking behavior. By contrast, age at menarche was not
associated with any of the three norm violation measures for the Slovakian girls. All
correlations were near zero. In tests of these differences using r-to-z transformations,
each of the correlations in the Swedish cohort was significantly stronger than the
corresponding correlation in the Slovakian cohort, Z = 3.27, p < .001, Z = 3.60, p <
.001, and Z = 3.72, p < .001 for sexual maturation, intercourse, and steady boyfriend,
respectively. Thus, these results are consistent with the assumption that Swedish
culture amplifies and Slovakian culture inhibits the link between girls’ early
maturation and problem behavior.
PAPER I
Table 2.3.
Percentages of early and later maturing Swedish and Slovakian girls reporting
romantic relationships with boys.
25
26
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
Table 2.4.
Correlations between pubertal timing and different aspects of norm violations for
Swedish and Slovakian girls.
Evaluations
Behavioral intentions
Behavior
* p < .05
** p < .01
Sweden
Slovakia
Z
-.23***
-.23***
-.24***
-.02
-.03
-.03
3.72 ***
3.60 ***
3.72 ***
*** p < .001
Links between romantic relationships with boys and norm violations in both
cultures. If socialization by older, more normbreaking peers, particularly boys,
explains why early pubertal maturation is linked to normbreaking, then romantic
relationships with boys should have about the same implications for normbreaking
behavior in Swedish and Slovakian societies. To test this prediction, we correlated
three measures of romantic relationships—sexual maturation, intercourse, and having
a steady boyfriend—with three measures of normbreaking—evaluations, behavioral
intentions, and normbreaking behavior. Table 2.5 shows these correlations.
Table 2.5.
Correlations between different aspects of romantic relationships with boys and norm
violations for Swedish and Slovakian girls.
________________________________________________________________________
Sweden
Slovakia
Sexual Intercourse
Steady
Sexual
Intercourse
Steady
maturation
boyfriend
maturation
boyfriend
________________________________________________________________________
Evaluations .21***
.19***
.25***
.23***
.24***
.17***
Behavioral .26***
intentions
.28***
.30***
.27***
.28***
.22***
Behavior
.36***
.43***
.37***
.32***
.38***
.26***
_________________________________________________________________________
Note: The Swedish and Slovakian correlations differ for steady boyfriend and normbreaking
behavior Z = 2.16, p < .05; All other Zs < 1.52, ps > .10; ** p < .01 *** p < .001
All three measures of romantic relationships were significantly correlated with the
three measures of normbreaking for girls in both cohorts, particularly with girls’ own
normbreaking behavior. Furthermore, in tests of the differences, 17 out of 18 did not
differ, all Zs < 1.52, ps > .10. In the only case where the correlations differed
significantly (Z = 2.16, p < .05 for the correlation between steady dating and
normbreaking), both correlations were statistically significant. Thus, even though
relationships with boys were less common among Slovakian girls than among
Swedish girls, they had the same implications in both cultures for normbreaking
behavior, which is consistent with the peer socialization hypothesis. It appears, then,
26
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
27
Discussion
In this study, we set out to test the peer socialization hypothesis by selecting cultural
contexts in which peer socialization should have a good chance of working (Sweden)
and a poor chance of working (Slovakia). We showed, first, that the conditions
favorable to peer socialization have apparently existed over at least two generations in
Sweden. Then, we showed that the same mechanisms do not work in Slovakia. When
girls’ contacts with boys are restricted, as they purportedly are in Slovakia, early
puberty is not linked to normbreaking behavior. When contacts with boys do occur in
Slovakia, however, they are linked to normbreaking just as in Sweden. Taken
together, these results lend support to the peer socialization hypothesis and the
assertion that the processes described in the hypothesis can be amplified or inhibited
on the societal or cultural level.
In this study, Slovakian society was chosen as the contrast to Swedish society
because of its conservative attitudes toward adolescent sexuality, particularly for girls.
While we do not know exactly how the differences we found might generalize to
other countries, we do know that permissiveness toward sexuality among adolescent
girls in Sweden is higher than in many other Western countries. Contrary to what is
often found for North American populations (Graber & Brooks-Gunn, 2002), the
point prevalence of sexual intercourse is typically higher for Swedish girls than for
boys throughout adolescence. In a Swedish study conducted in 2001 (Kerr, Stattin, &
Kiesner, 2007), similar percentages of 13-year-old girls and boys reported having had
intercourse, but for all other ages up to 17 years, the percentage was higher for girls
than boys, significantly so at most ages. Similar differences were found in 1970 for
Swedish 15-year-olds (Magnusson et al., 1975), in 1980 for 14 to 18-year-olds
(Lewin, 1982), and in 1990 for 17-year-olds (Edgard, 2002). Note that because girls
mature earlier than boys, one should expect girls to have their sexual debuts earlier
than boys if pubertal processes alone determined onset of sexual activity (Udry,
1990). These differences give one reason to expect similar differences between
Swedish samples and at least some North American samples.
One might be tempted to conclude from these findings that the liberal attitudes
toward sexuality in Swedish society are damaging to girls’ development. We should
mention, however, that although the acceptance of adolescent sexuality in Sweden
apparently allows peer socialization to work, it does not automatically translate into
high levels of problems such as teenage pregnancy, abortions, and sexually
transmitted diseases in Sweden compared with other countries. Rather, the opposite is
true. Sweden has, in international comparison, a low prevalence of abortions and
teenage pregnancies, and knowledge of contraceptive methods is good (Edgardh,
2002; Persson & Jarlbro, 1992). This has been attributed to obligatory sex education
in schools since the 1950s, free, on-demand abortions, low-priced contraceptives, and
a system of youth health clinics throughout the country that provide free contraceptive
counseling to adolescents.
It is commonly thought that it is early teenage sexuality that is at the heart of the
problem of teenage pregnancy, abortions, and sexually transmitted diseases, and
consequences like low education. Delaying its onset should solve the problem. This is
the message also in many empirical studies in the literature. For example, Crockett,
Bingham, Chopak, & Vicary (1996) in a study of sexuality among rural adolescents,
PAPER I
that the Swedish cultural context amplifies the link between early maturation and
normbreaking by making it easy for early maturing girls to have relationships with
boys, whereas the Slovakian cultural context inhibits this peer socialization process.
27
28
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
in the introduction discussed the problems of pregnancy abortions, and sexually
transmitted diseases among American females and stated that ”The timing of an
adolescent’s first sexual intercourse is a key variable affecting these negative
consequences… Adolescents who postpone first intercourse until their late teens or
early twenties may also merit special attention. These adolescents remain virgins after
the majority of their peers have initiated intercourse. Hence, studies of these youth
may help elucidate strategies for delaying first intercourse” (italics added, p. 90). In
an international comparison, though, a different picture emerges. Abortions are higher
and teenage pregnancy is substantially more common among US adolescents than
adolescents in comparable European countries, and in Canada (Darroch, Frost, &
Singh, 2001). This cannot be explained by timing of sexual activity, as the prevalence
of sexual intercourse at different ages does not vary much between the countries.
However, considerable differences appear for contraceptive use, with US adolescents
being those who use contraceptives least (Darroch et al., 2001). Thus, it does not
appear that delaying onset of intercourse will necessarily lower the high rates of
teenage pregnancy and abortion among US adolescents.
Apparently, it is important to consider the societal context of developing girls to
better understand the link between early puberty and normbreaking, as it seems as it is
only in cultures where there are liberal views on adolescent heterosexual involvement,
that we find this link. But what contexts at other, more proximal levels, do they play a
role in this link as well? According to our hypothesis, the peers’ of the early maturing
girls have a prominent role in the link between pubertal timing and normbreaking.
(We assume that older peers and boyfriends provide a context in which early maturing
girls are socialized into more advanced social behaviors.) Who are those peers?
Where do the early maturing girls meet their peers? Are the peers older and more
normbreaking as the peer socialization hypothesis assumes? These questions are the
focus of Chapter III.
28
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
29
A basic premise of the peer socialization hypothesis is that early maturing girls select
friends and romantic partners who are similar to them in maturity, namely older peers.
Earlier research has provided some support for this premise, showing that early
developing girls were more likely to report having older friends and sexual and
romantic relationships (Hayne, 2003; Stattin & Magnusson, 1990). Issues of context,
however, were not considered in these studies, and the contexts of peer relationships
should logically be expected to either facilitate or impede early developing girls’
attempts to affiliate with similar others. Specifically, the school context should make
it difficult for early maturing girls to affiliate with older peers, because most school
activities are segregated by age. Outside of school, however, early maturing girls
should be able to find contexts in which they can associate with older peers.
This reasoning about contexts implies two consequences for research. First, it
implies that the research design (including peers in or outside school) should
determine whether the findings support the peer socialization hypothesis. For
instance, in much research on adolescent peer associations, peers are defined as
school or classroom friends. With this design, one should not expect to find that early
developing girls associate with older peers. For instance, in a large North American
sample, early maturing girls had more deviant peers and more romantic relationships,
but not more older and male friends compared to on-time maturing girls (Haynie,
2003). These findings might have been influenced by the fact that girls could only
nominate peers who were at their schools (and not older peers who had left school or
were at other schools). When girls have reported on their friends in and outside of
school, early maturing girls have been shown to associate with older peers (Stattin &
Magnusson, 1990). Thus, to the extent that contexts are important, the research design
will determine what answers one can get. Second, this reasoning about contexts
implies that research designs including contexts can be used to test the basic premise
of the peer socialization hypothesis. If this basic premise of the peer socialization
hypothesis is correct, then early developing girls should have different peer networks
in school and outside of school, whereas on-time and late developing girls should not.
In school, where the context discourages affiliation with older peers, peer groups
should be made up largely of same-age peers for early and later developing girls alike.
Outside of school, the peer groups of early developing girls should include older
peers, whereas those of later developing girls should not.
Examining peer networks outside of school presents another challenge—getting
valid information about peers’ behaviors. The problem is that people see others,
particularly their friends, as more like them than they are, a phenomenon known as
the false consensus effect (Oliver, Bakker, Demerouti, & de Jong, 2005; Ross et al.,
1977; Wolfson, 2000). This has been demonstrated in research on adolescent peer
relationships. Youths report their peers’ problem behaviors as more similar to their
own than the peers’ own reports show them to be (Iannotti, Bush & Weinfurt, 1996).
Thus, the challenges for examining the peer networks of early maturing girls include
both assessing peers in and outside of school and obtaining reports of peers’ behaviors
that are independent of the girls, themselves.
In this chapter, we report results from a study in which we did this by assessing all
youths from grade 4 through grade 12 in an entire community. This is roughly all
youths from ages 10 through 18. Participants reported which peers they engaged with
in school and outside of school. Because we collected information from all youths in
the community, we can identify each girl’s peers, regardless of school, class, age, or
PAPER I
Chapter III: School versus free-time contexts
29
30
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
gender. With this design, all potentially important peers are included in the study, and
their behavior is not filtered through the eyes of the target girl. We use data from 14and 15-year-old girls to examine the basic premise of the peer socialization
hypothesis: early maturing girls will affiliate with older peers and boyfriends. Based
on the idea that free time contexts will facilitate affiliation with older peers and
boyfriends whereas school contexts will inhibit it, we predict that pubertal timing will
be related to the age and delinquency levels of the peers girls associate with
exclusively in free time, but not the peers they associate with in school. That is, the
earlier girls matured, the older and more normbreaking their free-time peers should
be. We examine this with peers’ reports of their own behavior. In addition, we predict
that early pubertal timing will be associated with performing more normbreaking acts
with the free-time peer group. We examine this with girls’ reports of activities with
free-time peers. Confirmation of these predictions will show support for the peer
socialization hypothesis and for facilitation of associations with older peers as one
possible mechanism of contextual amplification.
Method
The community and participants. The study took place in a Swedish
community of about 26,000 inhabitants, which was chosen because it met several
criteria. First, it has its own high school, so youths remained in the city school system
over the entire adolescent period. Second, it did not offer youths easy access to
neighboring cities, so it was reasonable to assume that most youths had their leisure
time peer associations inside the community. Third, it had a population of the
appropriate size so that all potentially influential peers could be included. The
community had 13 schools and 158 classes in the age range covered. We will use data
from 8th and 9th grade girls the first wave. Of the 334 girls enrolled in grades 8 and 9,
309 (93%) participated in the study.
Measures. Pubertal maturation was a measure of the age at menarche, the
same measure as used in the previous studies.
School and free-time peer groups. The participants were asked whether they
belonged to a group of at least three friends that hung out together at school. They
were asked to list the names of up to ten peers belonging to this group. Participants
were also asked if they belonged to a group that hung out together in free time, and
they listed up to ten people who belonged to this group.
To isolate the free-time friends, we constructed three peer groups. (a) School
peers, or the peers girls mentioned in their in-school groups but not in their free-time
groups; (b) Conjoint peers, or those who were mentioned in both the in-school and
free-time groups; and (c) Free-time peers, or those whom girls mentioned in their
free-time groups but not in their school groups. Hence, free-time peers were those
whom girls spent time with exclusively in free time.
Of the 309 girls, 274 reported in-school groups, free-time groups, or both. Of
these 274 girls, 217 (79%) had friends they associated with only at school, 155 (57%)
had friends they associated with both in and outside school, and 115 (42%) had
friends they associated with only during free-time. The early maturing girls did not
mention more peers either in or outside of school than the late maturing girls. The
correlations between pubertal timing, on one hand, and having in-school, conjoint,
and free-time peers were low and non-significant (all rs < .04). Neither was pubertal
timing related to numbers of peers in any of these groups. Hence, the numbers of
peers in different contexts does not differ between the early and the late maturing
girls, but the characteristics of peers in these different contexts might differ.
30
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
31
Results
Age and delinquency of in-school and free-time peer groups. The
early maturing girls are hypothesized to select older and more delinquent friends, but
the school setting should inhibit their freedom to do so. Thus, differences in peer
characteristics between early and late maturing girls are most likely to be true of the
peers the girls meet exclusively during free-time. To test this, we correlated pubertal
timing with age and delinquency information for the three peer groups. The results are
reported in Table 3.1. As shown in the table, pubertal timing was not related to age or
delinquency of school or conjoint peers, those whom girls spent time with in the agegraded school context. Pubertal timing was, however, related to age and delinquency
of free-time peers.
PAPER I
Age of peers. Peers in the school and free-time groups were identified through
class lists. We calculated their ages from information on the class lists about dates of
birth.
Delinquency of peers. We computed summary delinquency measures for each of
the peer groups from the individual peer group members’ self-reports of delinquency.
The delinquency measure was composed of 22 questions: “Have you taken things
from a store, stand, or shop without paying;” “Have you been caught by the police for
something you have done;” “Have you, on purpose, destroyed things such as
windows, street lights, telephone boots, benches, gardens, etc.;” “Have you taken
money from home that was not yours;” “Have you been part of painting graffiti, or
writing with markers or spray paint on, for example, a sidewalk;” “Have you
participated in breaking into a home, shop, stand, storage building or other buildings
with the intention of taking things;” “Have you stolen something from someone’s
pocket or bag;” “Have you bought or sold something that you knew or thought had
been stolen;” “Have you taken a bicycle without permission;” “Have you threatened
or forced someone to give you money, cigarettes, or anything else;” “Have you taken
part on a fight on the street in town;” “Have you carried a weapon (e.g., brass
knuckles, club, knife, switchblade, or some other weapon);” “Have you participated in
taking a car without permission;” “Have you taken a moped, motorcycle, or vespa
without permission;” “Have you been part of hitting someone so that you believe or
knew that he or she needed to be treated at the hospital;” “Have you intentionally hurt
someone with a knife, switchblade, brass knuckles, or some other weapon;” “Have
you taken part in threatening or forcing someone to do something that her or she
didn’t want to do;” Have you snuck out without paying (e.g., at movies, a café, on the
train or bus, or somewhere else,” “Have you taken part in stealing something from a
car;” “Have you drunk so much beer, liquor, or wine that you got drunk;” “Have you
smoked hashish (marijuana, cannabis);” and “Have you used any drugs other than
hashish (marijuana, cannabis)?” The alpha reliability for a scale summing these items
was .92.
Activities the last month with free-time peer groups. Participants were asked
what activities they had done with their free-time peers the last month. They answered
on a scale from (1) no to (3) yes, several times for the following 11 activities: Studied;
Watched TV or videos; Played hooky; Talked on the telephone/chatted; Shoplifted;
Used the computer, Internet, or played TV videogames; Drank alcohol until you were
drunk; Kept secrets from parents; Talked about things that are illegal; Went out
looking for boys; and Did other things, apart from shoplifting, for which one could get
caught by the police.
31
32
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
Table 3.1.
Correlations between pubertal timing and age, gender, and delinquency characteristics of
girls’ peer groups at school, at school and in free-time (conjoint), and in free-time (one-tail
tests).
_______________________________________________________________________
Peer characteristics
_________________________________________________
Age
Delinquency
_______________________________________________________________________
School peers
.05
.04
(173)a
(163)
Conjoint peers
.05
-.01
(111)
(107)
Free-time peers
.37 ***
.29 **
(73)
(97)
_______________________________________________________________________
a
n; ** p < .01; *** p < .001
Activities with free-time friends. We also examined the association
between age at menarche and the activities girls had done with their free-time peers.
The findings are reported in Table 3.2. As shown in the table, pubertal timing was
associated with the activities in this list that would be considered normbreaking. It
was significantly related to keeping secrets from parents and drinking alcohol to the
point of intoxication with free-time friends, and it was marginally significantly related
to being truant from school with friends. These results support the idea that early
maturing girls do more normbreaking activities with their free time friends than do
their later-matured counterparts.
Table 3.2
Correlations between age at menarche and activities the last month with the free-time peer
group (one-tail tests).
___________________________________________________
r
___________________________________________________
Studied
-.03
Watched TV/Videos
.06
Played hookey
.12 †
Talked on the telephone/chatted
.03
Shoplifted
.08
Used the computer, internet
.04
Drank alcohol until you were drunk
.17 **
Kept secrets from parents
.24 ***
Talked about things that are illegal
.09
Went out looking for boys
.02
Other things for which one gets caught
.05
____________________________________________________
† p < .10; * p < .05; ** p < .01; *** p < .001
32
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
33
A basic premise of the peer socialization hypothesis is that early maturing girls will
choose to associate with peers who are older than they and, by extension, more
delinquent. Although there is support for this in the literature (Stattin & Magnusson,
1990), there are also findings where early maturing girls do not have older peers
(Haynie, 2003). In this chapter, we hypothesized that age-graded contexts such as
school should interfere with this peer selection process, and non-age-graded, free-time
contexts should facilitate it. As predicted, age of puberty was not related to the ages or
delinquency levels of friends girls associated with in school, but it was related to the
ages and delinquency levels of friends girls associated with outside of school. These
results show support for a basic premise of the peer socialization hypothesis—that
going through puberty early spurs girls to seek out older peers who are higher on
problem behaviors than their same-age peers. The results also illustrate one possible
way in which contextual amplification might take place—by facilitating associations
with older, more problematic peers.
This study has more general implications for the study of peer influences on
problem behavior. Many studies in this area have dealt with peer groups in school
(e.g., Adler & Adler, 1998; Campbell, 1980; George & Hartmann, 1996; Giordano,
Cernkovich, & Pugh, 1986; Hartup, 1983). This has been based on findings that the
majority of youths’ friends are of the same age and sex and go to the same school
(Coleman, 1961). In fact, this is probably true. For instance, in one study of the link
between pubertal timing and delinquency among 7th to 12th grade girls, participants
nominated up to 10 friends, 80% of whom were found in school rosters (Haynie,
2003). The conclusion drawn was that the school peers satisfactory captured the
overall friendships network. The findings in the present study are to a certain extent
the same. Most of the girls mentioned peers in their own school. But this is not to say
that these peers were the most important for the girls’ behavior. In fact, the minority
of peers that these girls and other youths engage with outside of school might have
strong implications for their behaviors (see also Kerr, Stattin, & Keisner, 2007).
Studies of peer influence based on classroom peers tend to miss these potentially
important peers.
A limitation of this study is that we did not distinguish between different types of
free-time contexts. There are many different types of free-time contexts available to
adolescents, and some could facilitate associations with problematic peers, whereas
others could hinder them. In the next chapter, we examine this issue.
PAPER I
Discussion
33
34
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
Chapter IV: Leisure settings
In Chapter II we looked at societal conditions that should amplify or inhibit the peer
socialization effect. In Chapter III we examined characteristics of peers girls
associated with at school or during leisure time. In the present chapter, we look at
leisure time context that should amplify the peer socialization effect. If the peer
socialization hypothesis is correct, early pubertal timing should be linked to problem
behavior for girls who spend time in leisure settings that provide opportunities for
affiliation with more advanced or more problematic peer groups. In particular, a
leisure setting should facilitate the peer socialization effect if early maturing girls: (a)
can meet boys and peers who are older than they are and (b) are likely to meet
adolescents who are high on normbreaking behavior.
In this study we examine one example of such a setting: neighborhood youth
recreation centers. These centers allow young adolescent girls to socialize with older
girls and boys, and the youths who spend a lot of time there tend to be higher on
antisocial behavior than youths who do not (Mahoney & Stattin, 2000; Mahoney et
al., 2001). If the peer socialization hypothesis is correct, then early pubertal
maturation should be more strongly linked to normbreaking for girls who attend these
centers than for those who do not.
Youth centers exist in most European and North American communities to
provide recreation opportunities for youths who might otherwise be loitering on the
streets, drinking, and getting into trouble. Adults are present in these centers, but
seldom direct the activities. The aim is to bring at-risk youths into a semi-supervised
situation. In Sweden, local youth recreation centers are available in most communities
to adolescents 13 years and older. Typically, they are accessible every evening of the
week, opening around dinnertime and closing as late as 11:30 p.m. They usually offer
pool, ping-pong, video games, darts, TV, music, and coffee. Adults are present at
these centers, but they do not direct the youths’ activities or place any demands upon
them. The general idea is that spending time at a recreation center is less risky and
potentially more constructive than the alternative ways that youths who are not
involved in organized activities could spend their free time.
Even though these centers have been organized with the best of intentions,
research suggests that in practice they might not be beneficial. In studies of Swedish
centers, youths who attended the centers regularly were more antisocial, had more
antisocial peers, had more conflicted relationships with parents, and had parents who
knew less about their activities than those who did not attend the centers (Mahoney &
Stattin, 2000). Boys with multiple problems tended to go to the centers, and even after
controlling for earlier levels their problems increased in adolescence and into
adulthood more than those of boys who did not go to the centers (Mahoney et al.,
2001). Thus, these findings suggest that poorly adjusted youths are attracted to the
youth centers, and after they begin participating, they become socialized into even
more problem behavior.
The youth centers should facilitate the processes outlined in the peer socialization
hypothesis. They are not age-graded, so early maturing girls can associate with older
peers who match their biological maturity. These peers will be higher on average in
normbreaking just by virtue of being older. In addition, however, youths at the centers
tend to be higher in normbreaking than others their age, thus suggesting that a peer
socialization effect should be intensified. In a study of youth center attendance in
girls, the results suggested that girls who frequently visited the centers were likely to
meet antisocial boys and be influenced by them (Stattin, Kerr, Mahoney, Persson, &
34
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
35
Method
Participants. The 1970 cohort. These girls were the main group followed
longitudinally from age 10 to age 43 (Magnusson et al., 1975). They are described in
more detail in Chapter II.
The 1998 cohort. This group of girls was part of a short-term longitudinal study.
Fourteen-year-old youths in Örebro, a Swedish city of 120,000, were followed over
18 months, from 1998 to 2000 (Kerr & Stattin, 2000; Stattin & Kerr, 2000). Students
in all 8th grade classes in the city (N = 1279) composed the target sample (48% boys;
52% girls) in the first wave of the longitudinal study. They took part in the study
unless their parents returned a form stating that they did not want their child to
participate (12 parents returned this form). Of the 1279 students, 1186 (93%) with a
mean age of 14.42, were present on the day of the data collection at the first wave and
answered the questionnaires.
Analyses of the 1970 cohort cover ages 13 to age 15: peer involvement and center
attendance were measured when the girls were 13 years old (grade 6), and information
about girls’ romantic relationships with boys, problem behavior, and pubertal timing
was gathered when they were 15 years (grade 8). For the 1998 cohort, analyses cover
age 14 to age 16 (over 18 months).
Measures. Two of the five measures were the same for the 1970 and the
1998 cohorts.
PAPER I
Magnusson, 2005). Pubertal maturation was not considered in that study, however,
and this should be particularly true for early maturing girls. Thus, if the peer
socialization hypothesis is correct, then early maturing girls who spend a lot of time at
neighborhood youth recreation centers should be higher on normbreaking than those
who do not.
Assuming that the results appear as predicted, however, one might question
whether the problematic adjustment of the early developed girls is something that is
circumscribed to the mid-adolescent years or has long-term implications for adult
social adjustment. It should be mentioned, first, that higher registered criminality in
adulthood was found for boys who attended youth recreation centers at age 13 than
for those who did not (Mahoney et al., 2001). This study controlled for relevant
behaviors at age 10, before the boys started to visit the centers. So, attending the
centers can have long-term negative implications for boys. The question is whether
the same is true for girls. Some support for sustained social adjustment problems
among the early maturing girls has been reported earlier (Stattin & Magnusson, 1990).
A somewhat higher proportion of the earliest maturing girls than of the late maturing
girls was registered by the police for some offence up to age 30 (8.2% of the girls who
had their first menstruation before 11 years were registered for some offence up to age
30, versus 4.1% of the girls who had their first menstruation at age 13 or later). This
was a rough comparison, however, disregarding other relevant factors in adolescence
that might influence future offending.
We present findings from studies of two cohorts. First, we predict problem
behavior at age 15 from information about heterosexual relations and attending youth
recreation centers two years earlier, at age 13, in a long-term longitudinal study—the
1970 Cohort. We cross-validate this study using information from a short-term
longitudinal study—the 1998 Cohort. Then, using adult information from the 1970
Cohort we predict future registered criminality from center attendance and early
heterosexual relationships.
35
36
Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
Youth center participation. In both cohorts, participants were asked whether
they usually attended the youth centers in the town. For the 1970 cohort, the response
options involved a 5-point scale ranging from no participation to participation every
evening of the week. For the 1998 cohort, the response options were (1) No or (2) Yes.
If they answered yes, they were asked how many evenings per week they typically
spent at a youth center. Note that an adolescent has to be 13 years of age to have
entrance to these centers.
Pubertal timing was measured in the same way as described in previous studies.
Romantic relationships with boys. In both cohorts, three items measured how
advanced the girls’ relationships to boys were. These items were described in Chapter
II. They covered perceived sexual maturation, having had sexual intercourse, and
having a steady boyfriend.
Problem behavior. In the 1970 cohort, problem behavior was a scale aggregated
from eleven questions about shoplifting, alcohol abuse, signature forgery, truancy,
using hashish, vandalism, evading payment in various places, hitting somebody,
harassment, and loitering in town in the evening (Magnusson et al., 1975). It was the
same scale as described in Chapter II. In the 1998 cohort, the scale included more
serious delinquency items. Youths answered 15 questions about whether they had
engaged in certain behaviors during the past year. The response scale was a 5-point
scale ranging from “never” (1) to “more than 10 times” (5). The questions were
about: shoplifting; being caught by the police for something they had done;
vandalizing public or private property; taking money from home; creating graffiti;
breaking into a building; stealing from someone’s pocket or bag; buying or selling
stolen goods; stealing a bike; being in a physical fight in public; carrying a weapon;
stealing a car; stealing a moped or motorcycle; using marijuana or hashish; and using
other drugs. The alpha reliability was .72.
Peer involvement. In the 1970 cohort, peer involvement was a composite
measure of three questions “How many peers do you have that you are together with
really often after school?” “How many evenings per week do you usually spend with
your peers?” and “How many evenings per week do you usually spend at home?”
(reversed). The inter-item correlation for these three measures was acceptable, .34, but
the alpha was low, .64. In the 1998 cohort, youths reported how many nights per week
they usually socialized with friends. The response options ranged from one to seven
days per week.
Registered offenses. For the 1970 cohort, information about criminal offences
was obtained from official registers. For underage offending (below age 15),
information about juvenile offending was obtained from local welfare authorities.
Information about registered offences between age 15 and age 30 were obtained from
police registers. Only 3 of the girls had underage offending. At age 30, 38 (7.5%) of
the 510 girls with data on menarcheal age were registered by police for one or several
criminal offences.
Stattin, Magnusson, & Reichel (1989) compared rates of criminality in Örebro
with other major cities in Sweden, and with the nation as a whole, across the years
1965 - 1980. Overall, the crime rate in Örebro was higher than the Swedish national
average during this time; however, the gap closed over the years. This was the result
of a rise in narcotic and fraud offences that increased for the nation as a whole, but not
in Örebro. A city-based comparison showed that Örebro had comparable rates of
crime to similar-sized cities in Sweden (50,000-100,000 inhabitants), and moderately
lower rates of crime compared to Sweden’s largest cities (i.e., Stockholm, Göteborg,
Malmö).
36
Female
FemalePubertal
PubertalTiming
Timingand
andProblem
ProblemBehavior
Behavior
37
Results
Descriptive Information. Bivariate correlations for the main measures used
in the study appear in Table 4.1, with correlations for the 1970 cohort below the
diagonal and those for the 1998 cohort above the diagonal. Problem behavior at age
15 and delinquency at age 16 were both related to frequency of youth center
attendance two years earlier and to pubertal timing. The strongest correlations in both
cohorts were between problem behavior and romantic involvement and between
center attendance and peer involvement. Center attendance was unrelated to pubertal
timing.
Tests of Predictions from the Peer Socialization Hypothesis. We
expected the youth center context to amplify the peer socialization effect by
facilitating contacts between early maturing girls and peers who are older and, on
average, higher on delinquency than those who do not frequent the youth centers.
Thus, we predicted that delinquency or problem behavior would be a function of
pubertal timing, center attendance, and either peer or heterosexual involvement. To
examine this prediction, we tested multiple regression models in each cohort
predicting problem behavior (in the 1970 cohort) and delinquency (in the 1998
cohort) from pubertal timing, center attendance, and peer/heterosexual involvement
and all of the two-way and three-way interactions. We predicted three-way
interactions. The results are shown in Table 4.2, with the models including
heterosexual involvement in the upper part of the table and models including peer
involvement in the lower part.
PAPER I
Analyses. We used multiple regression analyses, centering the measures
before calculating the interaction terms.
37
Table 4.1.
Correlations among the main measures for the 1970 cohort (below the diagonal) and the 1998 cohort (above the diagonal).
______________________________________________________________________________________________________________
Problem Behaviora/
Peer
Romantic
Center
Pubertal
involvement
relationships
attendance
timing
Delinquencyb
______________________________________________________________________________________________________________
Problem behaviora/Delinquencyb
—
.35 ***
.36 ***
.32 ***
-.24 ***
(627)
(620)
(514)
(621)c
Peer involvement
.25 ***
—
.32 ***
.40 ***
-.03
(429)
(616)
(608)
(503)
Romantic relationships
.47 ***
.33 ***
—
.28 ***
-.19 ***
(517)
(427)
(612)
(507)
Center attendance
.24 ***
.34 ***
.19 ***
—
-.05
(433)
(518)
(431)
(504)
Pubertal timing
-.24 ***
-.14 **
-.34 ***
-.08
—
(510)
(422)
(508)
(426)
______________________________________________________________________________________________________________
a
1970 cohort; b1998 cohort; cn; *p < .05; **p < .01; ***p < .001
38
Female Pubertal Timing and Problem Behavior
PAPER I
Table 4.2.
Simultaneous regression models predicting problem behavior (1970 cohort) or delinquency (1998 cohort) at age 15.
___________________________________________________________________________________________
1970 cohort
1998 cohort
______________________
______________________
beta
Model R2
beta
Model R2
___________________________________________________________________________________________
Romantic relationships
Romantic relationships with boys
.36 ***
.11 *
Center attendance
.08 ns
.23 ***
Pubertal timing
-.06 ns
-.18 ***
Romantic x Center
-.04 ns
-.04 ns
Romantic x Puberty
-.08 ns
-.15 **
Center x Puberty
-.11 *
-.05 ns
Romantic x Center x Puberty
-.20 **
.34 ***
-.14 **
.22 ***
Peer involvement
Peer involvement
.19 ***
.30 ***
Center attendance
.10 ns
.10 ns
Pubertal timing
-.18 ***
-.12 **
Peer x Center
.08 ns
.04 ns
Peer x Puberty
-.03 ns
-.32 ***
Center x Puberty
-.21 ***
.10 ns
Peer x Center x Puberty
-.04 ns
.16 ***
-.25 ***
.32 ***
___________________________________________________________________________________________
*p .05; **p .01; ***p .001
Female Pubertal Timing and Problem Behavior
39
40 Female Pubertal Timing and Problem Behavior
Female Pubertal Timing and Problem Behavior
0.7
Problem behavior
0.5
Girls who seldom
visit youth centers
Pubertal
timing:
0.3
Early
Late
0.1
-0.1
-0.3
-0.5
Heterosexual involvement
Little
Figure 4:1a.
Much
40
1
Female Pubertal Timing and Problem Behavior
41
PAPER I
0.7
Problem behavior
0.5
Girls who often
visit youth centers
Pubertal
timing:
0.3
Early
Late
0.1
-0.1
-0.3
-0.5
Heterosexual involvement
Little
Much
Figure 4.1b.
Figure 4.1. Three-way interaction between heterosexual involvement, pubertal timing, and youth
center attendance in predicting problem behavior in the 1970 cohort.
Heterosexual involvement. As seen in the upper part of Table 4.2, for heterosexual
involvement, the three-way interactions were significant in both cohorts. We plotted these
interactions by solving the regression equation for different points. The plots appear in Figures
4.1 and 4.2. As shown in Figure 4.1, for the 1970 cohort heterosexual involvement was, in
general, related to higher levels of problem behavior, but this was especially so for early maturing
girls who frequently visited the youth centers. As shown in Figure 4.2, for the 1998 cohort
heterosexual involvement was related to delinquency for early maturing girls, in general, but
more so for girls who frequently visited the youth centers than for those who did not. In sum,
these cohorts were separated by almost three decades, and somewhat different measures of
problem behavior were used in the two studies. Nevertheless, in both cohorts the findings are as
predicted by the peer socialization hypothesis and by the idea that the youth recreation centers are
a context that amplifies the peer socialization effect.
2
Female
Pubertal Timing and Problem Behavior
42
0,9
Delinquency
0,7
Girls seldom at
youth centers
0,5
Early
Late
0,3
0,1
-0,1
-0,3
-0,5
Heterosexual relations
Little
Much
Figure 4:2a.
Pubertal
timing:
3
Female Pubertal Timing and Problem Behavior
43
PAPER I
0,9
Delinquency
0,7
Girls often at youth
centers
Pubertal
timing:
0,5
Early
Late
0,3
0,1
-0,1
-0,3
-0,5
Heterosexual relations
Little
Much
Figure 4:2b.
Figure 4.2. Three-way interaction between heterosexual involvement, pubertal timing, and youth
center attendance in predicting delinquency in the 1998 cohort.
4
Female
Pubertal Timing and Problem Behavior
44
1
0,8
Delinquency
0,6
0,4
0,2
Girls who seldom
visit youth centers
Pubertal
timing:
Early
Late
0
-0,2
-0,4
-0,6
-0,8
-1
Peer involvement
Little
Figure 4:3a.
Much
5
Female Pubertal Timing and Problem Behavior
45
PAPER I
1
Delinquency
0,8
0,6
0,4
Girls who often
visit youth centers
Pubertal
timing:
0,2
0
-0,2
Early
Late
-0,4
-0,6
-0,8
-1
Peer involvement
Little
Much
Figure 4:3b.
Figure 4.3. Three-way interaction between peer involvement, pubertal timing, and youth center
attendance in predicting delinquency in the 1998 cohort.
Peer involvement. As seen in the lower part of Table 4.2, the predicted three-way interaction
between peer involvement, center attendance, and pubertal timing appeared in one of the two
cohorts. In the 1998 cohort, the three-way interaction was significant, and a plot of the interaction
appears in Figure 4.3. As shown in the figure, peer involvement was related to delinquency for
girls who matured early, but not for those who matured later, and frequent youth center
attendance seemed to amplify this effect. Thus, the findings are consistent with predictions from
the peer socialization hypothesis and from the idea that the youth centers are a context that can
amplify the peer socialization effect. In the 1970 cohort, the three-way interaction was not
significant, but the two-way, Center attendance x Pubertal timing interaction was significant. A
plot of that interaction appears in Figure 4.4. From the figure, it is evident that peer involvement
was more strongly related to problem behavior for girls who matured early than for those who
matured later. Thus, in the 1998 cohort there is evidence for the peer socialization hypothesis and
for contextual amplification in the youth-center context. In the 1970 cohort there is support for
6
Female
Pubertal Timing and Problem Behavior
46
the peer socialization hypothesis, but not for the youth centers as a context that can amplify the
effect.
Problem behavior
0,5
0,4
0,3
0,2
0,1
0
Pubertal
timing:
Early
Late
-0,1
-0,2
-0,3
-0,4
-0,5
Center attendance
Little
Much
Figure 4.4. Two-way interaction between center attendance and pubertal timing in predicting
problem behavior in the 1970 cohort.
Registered offenses as the measure of delinquency. In the 1970 cohort, data on all
officially registered offenses up to age 30 existed, and we used this to examine the joint effects of
heterosexual involvement, youth center attendance, and early pubertal timing. Because of the
very low frequency of registered offending among girls, we dichotomized all measures and used
nonparametric statistics. We dichotomized pubertal timing as early (before age 12) versus later
(at age 12 or later); youth center attendance as attendance versus nonattendance; and heterosexual
involvement as having had sexual intercourse by age 14:10 versus not. We conducted separate
analysis of registered criminality within early and late maturers, dividing each group into two: (a)
those who did not attend youth centers and had not had sexual intercourse with (b) those who
attended the centers, had sexual experiences, or both. As reported in Table 4.3, among early
maturers, three times as many of the latter group (17.9%) than the former (5.7%) was registered
7
Female Pubertal Timing and Problem Behavior
47
% with criminal records up to age 30
35
30
25
20
Percentage of all girls in the
population who were
registered for a crime (7.6%)
15
10
5
0
sexcenter-
sexcenter+
sex+
center-
sex+
center+
Later maturing girls
sexcenter-
sexcenter+
sex+
center-
sex+
center+
Early maturing girls
Figure 4.5. Percentage of girls in the 1970 cohort with criminal records up to age 30 by pubertal
timing, early sexual intercourse, and youth center attendance.
PAPER I
for an offence up to age 30. Among late maturers, however, about the same proportions of the
two groups were registered for offenses (5.4% versus 5.7%).
For illustrative purposes, we present in Figure 4.5 registered offenses for the early and later
developing groups, each broken down into four subgroups according to sexual intercourse and
center attendance. These results must be interpreted with caution, because of the low numbers of
participants in each group. Nonetheless, the figure illustrates that the risk associated with the
combination of center attendance and heterosexual involvement is especially relevant for early
maturing girls. This is in accordance with our hypothesis: the risk for future criminal behavior
should be particularly concentrated to the early maturing girls who spend time in contexts with an
overrepresentation of antisocial peers and who become involved with boys.
48
8
Female
Pubertal Timing and Problem Behavior
Discussion
In the previous chapter, we linked early maturation with having older and more delinquent freetime peers. In this chapter, we focused on one free-time context that is known to attract
delinquent youths—both boys and girls. Our results were largely supportive of the prediction that
early maturation would be most strongly linked to delinquency for girls who spent time in this
context and had much involvement with peers and boys. Thus, the results were consistent with
the peer socialization explanation of the link between girls’ early maturation and problem
behavior, and also with the assumption that some contexts can facilitate peer socialization and in
so doing amplify the peer socialization effect.
The major limitation of this study is that we cannot say whether girls developed their
relationships with peers and boys at the youth centers. The questions about peer and romantic
relationships were asked independent of center attendance. We have assumed that frequent
attendance at the youth centers makes it likely that a girl’s peers and boyfriend would also be
spending time at the same center. This is probably a fairly safe assumption, but this conceptual
limitation should be recognized.
Despite this limitation, we believe there are several reasons for confidence in these results. It
is difficult to predict and find three-way interactions, and the fact that this predicted result
appeared in three analyses out of four argues for the robustness of the phenomenon. A further
argument for the robustness of the phenomenon can be made from the fact that similar results
appear in datasets collected almost 30 years apart in different samples and using measures of
problem behavior with different degrees of seriousness. That the same pattern of results found for
self-reported problem behavior can be seen in registered delinquency gives us further confidence
that the phenomenon is robust.
There is, however, one alternative explanation for these results that we have not yet
considered—that individual characteristics such as impulsivity or risk taking might underlie early
maturation, heterosexual involvement, and problem behavior. Girls with risky personality traits
such as thrill-seeking and impulsivity might be attracted to the youth centers and might also get
involved with boys and problem peers. These same traits have been linked to delinquency
(White, Moffitt, Caspi, Bartusch, Neddles & Stouthammer-Loeber, 1994; Cooper, Wood, Orcutt
& Albino, 2003; Leas & Mellor, 2000); consequently, they could explain the connection between
attending the youth centers and normbreaking. One study has looked at this for girls in general
(Persson, Kerr, & Stattin, 2004). The results showed that these characteristics partly explained
youth center attendance and involvement with boys, but they did not explain why those who
attended the centers and were involved with boys were highest on normbreaking. However, girls’
pubertal timing was not considered, and that leaves open the possibility that early maturing girls
with these traits go to the youth centers as a way to meet older peers and boys who offer
excitement. For these girls, personality traits such as thrill-seeking and impulsivity might be a
more important factor than peer socialization. In the next chapter, we test the idea about preexisting differences in risky personality traits with analyses of two different samples.
9
Female Pubertal Timing and Problem Behavior
49
The timing of girls’ first menstruation is multiply determined. Genetic factors, weight and body
fat, amount of daily exercise, intensive athletic activity, protein intake, illness, and environmental
stress in childhood, have all been linked to age at menarche (Burt et al., 2006; Graber et al.,
1995). This raises the question whether the link between pubertal timing and problem behavior
that we have documented here could in fact be explained by something that existed before
puberty rather than being something instigated by peer processes and contextual conditions in
adolescence.
Earlier studies have addressed this issue in a couple of ways. Some have examined whether
the same genes influence timing of puberty and problem behavior. One study reported that the
link between father abandonment and early pubertal timing and behavioral problems is a function
of the androgen receptor gene (Comings, Muhleman, Johnson, & MacMurray, 2002). Others have
found little support for a common genetic influence (Burt et al., 2006; Jorm et al., 2004), and
have shown that the genetic influence on behavioral problems is highest among on-time maturers
and lowest among early and late maturers (Burt et al., 2006). At least three earlier studies looked
at conduct problems as possible predictors of age at menarche (Caspi & Moffitt, 1991; Graber et
al., 1995; Moffitt et al., 1992). Neither study found significant results, suggesting that the link
found in the literature between pubertal timing and problem behavior in mid-adolescence is
mainly due to conditions operating in adolescence, not earlier in childhood. However, in another
study, the authors suggested that puberty accentuates pre-existing differences between early and
late maturing girls (Caspi & Moffitt, 1991). The results of this study showed that the girls whose
externalizing problems increased most from age 13 to 15 were the early matured girls with high
externalizing problems as early as age 9. Notably, however, the results also showed that whether
the girls were high or low on behavioral problems, the early maturing girls scored higher on
behavior problems at ages 13 and 15 than the late maturing girls, and the pubertal timing effect
was considerably stronger at age 15 than at age 13. Thus, the there is little evidence that preexisting problems underlie both puberty and problem behavior, but some evidence that problem
behavior is intensified by puberty. So far only limited numbers of problem indications have been
examined pre-menarcheal, so to be able to understand whether early problem behaviors predict
pubertal onset and mid-adolescent problem behavior, a range of childhood problem measures is
needed. This is the aim of the present study.
In this study, we present two sets of analyses. First, we focus on the 1970 cohort from
Chapters II and IV (Magnusson et al., 1975), and we examine whether potential differences in
normbreaking behavior in mid-adolescence (age 15) between early and later maturing girls can be
traced back to individual differences in problem behavior at age 10, when most of the girls were
pre-pubertal. We study a diverse range of factors linked to late childhood peer relationships and
social and personal adjustment. Measures of adjustment are taken from self reports, parent and
teacher reports, and register information. In the study we also examine the accentuation
hypothesis (Caspi & Moffitt, 1991). If social adjustment, peer relationships measures, or both
differ between early and later-developed girls at age 10, they should differ even more as time
passes. For these reasons we have selected relevant comparison measures which have been
assessed recurrently in the longitudinal project. For both the age-10 and age-13 measures, we also
look at links to pubertal timing. The peer socialization hypothesis links girls’ behavior problems
to romantic relationships with boys. Therefore, in the present study we examine whether
normbreaking behavior and heterosexual involvement could be explained by pre-pubertal factors.
PAPER I
Chapter V: Pre-existing differences
50
10
Female
Pubertal Timing and Problem Behavior
In the second set of analyses, we revisit the 1998 cohort and analyses from Chapter IV. We
examine whether the three-way interactions reported in Chapter IV for the 1998 sample still hold
after controlling for personal characteristics that are linked to both early maturation and problem
behavior. Because these two sets of analyses are quite different, we describe their methods and
results separately.
Accentuation of Preexisting Differences
Method
Participants. The participants in the first set of analyses were the 1970 cohort used in
Chapters II and IV (Magnusson et al., 1975; Magnusson, 1988). The sample was recruited in
1965, when the youths were 10 years old, and the target sample consisted of all third-graders in
the town of Örebro, Sweden. Over 90% participated and were followed up longitudinally.
Extensive information has been collected from different sources: the children, their teachers and
parents, and official registers (see Magnusson et al., 1975). A second data collection was done
when the children were 13 years old. At this time, 85% (423/498) of the girls in the sample
completed an extensive school survey. The 75 girls who did not respond to the age-13 survey had
either moved out of the area, or were not able to complete the survey during the testing days. A
comparison of these 75 girls with the 423 who participated revealed no significant differences on
aggression, hyperactivity, achievement, peer preference, or criminal offending (Mahoney et al.,
2001).
Measures. Social Behavior. Information about behavioral problems was determined
from teacher-ratings and peer nominations when the participants were 10 and 13 years of age.
Data included teacher-ratings concerning aggressiveness, concentration problems, motor
restlessness, disharmony, and school fatigue. Teacher-ratings were made on a 7-point scale,
where higher values indicate more of the rated construct (Magnusson et al., 1975). High stability
over ages 19 to 13 has been reported previously (Backteman & Magnusson, 1978). It should be
noted that in most cases the teachers who made these ratings had had the opportunity to observe
the girls over three years. The teachers who had been the head teachers from age 7 to 10 made the
age 10 ratings, and the teachers who had been the head teachers from age 10 to 13 made the age
13 ratings. In almost all cases, the teachers who made the ratings at age 10 were different from
the teachers who did the ratings three years later.
Intelligence. The Westrin Intelligence Test, WIT III, was given to the participants at age
13 (Westrin, 1967). The test contains four subtests measuring inductive ability, verbal
comprehension, deduction, and spatial ability. The manual reports a split-half reliability of .93.
Grades. Grades in Swedish and Mathematics were available for the girls at ages 10 and
13. We aggregated the grades over the two ages.
Popularity in the class. To assess popularity at ages 10 and 13, students were asked to
imagine that they were being transferred to another classroom, but that not all of their classmates
could go to the new room (see Magnusson et al., 1975). They were then asked to rank all of their
classmates in the order in which they would like them to move. The students were asked to rank
girls and boys separately. The rankings were then combined across all informants and
standardized to form a continuum of popularity. A constant of 3 was added to avoid negative
scores. Higher scores on this measure reflect a greater number of peer nominations (labeled
popularity).
11
Female Pubertal Timing and Problem Behavior
51
Results
As reported in earlier chapters, there was a significant bivariate correlation between pubertal
timing and normbreaking (r = -.24, p < .001) and a significant correlation between pubertal
timing and romantic relationships with boys (r = .34, p < .001).
The differences in normbreaking at around age 15 for the early- and later matured girls might
be due to conditions that existed much earlier, rendering pubertal timing a spurious correlate.
Hence, we started by asking if pubertal timing is associated with problematic adjustment before
puberty (age 10). As reported in Table 5.1, with one exception we found little evidence that
individual differences in pubertal timing were related to earlier adjustment problems. The
exception was teacher reported motor restlessness. There was a small but significant correlation
suggesting that the earlier girls developed, the more motor restless they were at age 10. To the
extent that this signals more hyperactive behavior among early maturing girls, we would also
expect concentration difficulties to be significantly associated with pubertal timing. They were
not. However, at age 13, pubertal timing was significantly linked to teacher-rated motor
restlessness (r = -.20, p < .001) and concentration problems (r = -.11, p <.05).
PAPER I
Perceived popularity. The participants were also asked about how they, themselves, would
be scored according to nominations of their peers in the class. We used this as a measure of
perceived popularity.
Peer relationships. Four items made up a scale measuring poor peer relationships at school
at age 10; “Do you enjoy the peers in your class;” “Do you have fun with peers during the
breaks;” “Can you play with those you want to play with during the breaks;” and “How many of
your classmates do you like?” The alpha reliability for the 4-item scale was .65 (inter-item
correlation was .31). Six items made up a similar measure of poor peer relationships at school at
age 13: “Do you have a really good friend in your class;” “Do you enjoy your classmates;” “Do
you have a good time during the breaks;” “Do peers pick on you during the breaks;” “How many
of your classmates do you like;” and “Do you get along well with your classmates?” The alpha
reliability for the age-13 measure was .78 (inter-item correlation was .37).
Parental concerns. When the girls were 10 years of age, their parents were asked if they had
any concerns in the following domains with respect to the child: pocket-money, TV-watching,
doing homework, curfew time, peer relationships, smoking, and alcohol or sniffing, and
disobedience. The response scale was a dichotomized measure: Yes or No. The alpha reliability
for the scale combining these eight domains was .66. A similar parental concern scales was made
up at age 13 with an alpha reliability of .74. It involved seven questions to parents about the
child’s pocket-money, TV-watching, doing homework, curfew time, peer relationships,
disobedience, and going to bed.
Pubertal timing. Puberty timing was the measure of age at menarche reported in the earlier
chapters.
Normbreaking. The measure of normbreaking behavior was used in Chapter IV. It consisted
of eight norm violating activities at home (ignoring parents’ prohibitions, staying out late without
permission), at school (cheating on an exam, being truant), and during leisure time (smoking
hashish, getting drunk, shoplifting, loitering in towns in evenings).
Heterosexual involvement was the same as in Chapter IV. The four items dealt with
perceived sexual maturation, perceived popularity among boys, sexual intercourse, and steady
boyfriend relationships.
52
12
Female
Pubertal Timing and Problem Behavior
Table 5.1.
Correlations between pubertal timing and various measures of poor peer relationships and social
and personal adjustment problems at ages 10 and 13.
_________________________________________________________________
Age 10 (n = 389)
Age 13 (n = 447)
r
r
____________________________________________________________________
Poor peer relationships
.05
-.02
Peer nominations (Peer)a
Perceived popularity (Self)
.07
-.04
Poor peer relationships (Self)
.07
.06
Social and personal adjustment problems
Good school adjustment
-.03
.01
Hours away from school
-.03
-.01
Parental concerns (Parent)
-.10
-.06
Grades (Register information)
.07
.02
Intelligence (Test)
.03
.06
School fatigue (Teacher)
-.03
-.06
Aggressiveness (Teacher)
-.09
-.07
Motor restlessness (Teacher)
-.12 *
-.20 ***
Concentration problems (Teacher) -.01
-.11 *
Disharmony (Teacher)
-.09
-.06
_________________________________________________________________
* p < .05; *** p < .001; ainformant
Do Pre-existing Problems Explain the Link between Early Maturation and
Mid-adolescent Problem Behaviors? With empirical evidence for late-childhood
differences in motor restlessness between early and later maturing girls, the question is whether
the associations between pubertal timing and romantic relationships with boys and between
pubertal timing and normbreaking can be explained by early motor restlessness. To examine this,
we used multiple regression analyses predicting normbreaking and romantic relationships from
pubertal timing, controlling for the measures that significantly differentiated the early from the
late maturing girls in late childhood: motor restlessness as reported by teachers at ages 10 and 13.
Then, to test the hypothesis that motor restlessness interacts with pubertal timing to produce
differences in normbreaking behavior and heterosexual relations later on, we added the
interaction between motor restlessness and pubertal timing.
13
Female Pubertal Timing and Problem Behavior
53
Age 10
Motor restlessness (teacher rating)
Pubertal timing (self-report)
Interaction: Motor restl x Puberty
Normbreaking behavior
Romantic relationships
with boys
Beta & sign
Beta & sign
.12 *
-.22 ***
.13 **
-.29 ***
-.02
-.00
.07
R2
Model
F(3,368) = 9.32, p < .001
Age 13
Motor restlessness (teacher rating)
.22 ***
Pubertal timing (self-report)
-.19 ***
Interaction: Motor restl x Puberty
-.03
R2
.10
Model
F(3,422)=15.56, p<.001
* p < .05
** p < .01
*** p < .001
.11
F(3,383) = 16,11, p < .001
.21 ***
-.27 ***
-.03
.15
F(3,442)=25.08, p<.001
As reported in Table 5.2, age-10 motor restlessness was a significant predictor of both age-15
normbreaking and age-15 romantic relationships with boys. Pubertal timing, though, was a
stronger predictor of normbreaking and romantic relationships. The interaction was not
significant. So, apparently motor restlessness at age 10 and pubertal timing are two independent
predictors of later normbreaking and heterosexual relationships. The findings are about the same
for age-13 motor restlessness. As seen in Table 5.2, motor restlessness at age 13 was a somewhat
better predictor of later normbreaking than was pubertal timing and a considerably stronger
predictor of romantic relationships than age-10 motor restlessness. Again, the interaction term
was non-significant. Thus, the link between pubertal timing and mid-adolescent normbreaking
and between pubertal timing and engagement in romantic relationships with boys cannot be
explained by problems that existed at ages 10 and 13. We have done the same analyses for a
measure of hyperactive behavior (summing teacher reported motor restlessness and concentration
difficulties), and the finding are about the same as reported here for motor restlessness.
Our results show that the early developing girls had some pre-existing problems, particularly
with respect to motor restlessness. Even though levels of this problem do not explain the link
between early development and later problem behavior, the possibility remains that puberty
accentuates motor restlessness over time (from age 10 to age 13) and this accentuation leads to
problem behavior in mid-adolescence. We examined this possibility.
As reported in Table 5.2, age-13 motor restlessness was a better predictor of later
normbreaking and heterosexual relationships than was age-10 motor restlessness. Is this change
in motor restlessness from age 10 to 13 due to pubertal timing, and predictive of future
normbreaking and heterosexual relationships. In order to answer this question, we regressed age13 motor restlessness on age-10 motor restlessness and saved the residual (expressing the change
in motor restlessness from age 10 to 13). We then predicted normbreaking at age 15 from age-10
PAPER I
Table 5.2.
Predicting mid-adolescent normbreaking behavior and romantic relationships with boys from
pubertal timing, motor restlessness, and the interaction between pubertal timing and motor
restlessness at age 10 and age 13, separately.
54
14
Female
Pubertal Timing and Problem Behavior
normbreaking, the residual (change in motor restlessness), and pubertal timing. To test the
accentuation hypothesis, we added an interaction term: the interaction between pubertal timing
and the residual change. We also predicted age-15 heterosexual relationships using the same
independent measures. The findings are reported in Table 5.3.
Table 5.3.
Predicting mid-adolescent normbreaking behavior and romantic relationships with boys from
pubertal timing, motor restlessness at age 10, the change in motor restlessness from age 10 to
age 13, and the interaction between pubertal timing and the change in motor restlessness from
age 10 to age 13.
_________________________________________________
beta
sign
_________________________________________________
Predicting age 15 normbreaking:1
Motor restlessness age 10
.11
.03
Change in motor restlessness
.18
<.001
from age 10 to age 13
Pubertal timing
-.19
<.001
Interaction between pubertal
-.04
n.s.
timing and change in motor restl.
Predicting age 15 heterosexual relations: 2
Motor restlessness age 10
.13
.006
Change in motor restlessness
.17
.001
from age 10 to age 13
Pubertal timing
-.27
<.001
Interaction between pubertal
-.03
n.s.
timing and change in motor restl.
________________________________________________
1
F(4,364) = 9.65, p < .001; R2 = .10; 2 F (4,379) = 15,52, p < .001; R2 = .14
As can be seen in Table 5.3, motor restlessness at age 13 was a significant predictor of later
normbreaking and heterosexual relationships. Motor restlessness change from age 10 to age 13.
Also, this change is a significant predictor of later normbreaking and heterosexual relationships.
But this change does not seem to interact with pubertal timing. Hence, we found no evidence for
the accentuation idea. Finally, pubertal maturation is a unique predictor of mid-adolescent
normbreaking over and above the other terms in the model. The results for a measure of
hyperactive behavior (summing motor restlessness and concentration difficulties at both ages 10
and 13) produced similar findings as reported in Table 5.3.
15
Female Pubertal Timing and Problem Behavior
55
Method
Participants. Participants were girls in the 1998 cohort from Chapter IV.
Measures. Thrill seeking was a subscale of the Youth Psychopathic traits Inventory
(YPI), an adolescent self-report measure of psychopathic personality traits (Andershed, Kerr,
Stattin, & Levander, 2002). The following 9 items, reported on a 4-point scale ranging from (1)
does not apply , to (4) apply precisely, are included: ”I get bored quickly by doing the same thing
over,” ”I get bored quickly and I like change,” ”I like to do things just for the thrill of it,” ”I like
to do exciting and dangerous things, even if it is forbidden or illegal,” ”I like people that are
involved in exciting and unexpected activities,” ”I have an unusually great need for change,”
”I’m drawn to places where exciting things happen,” ”I like having ’lively and active’
surroundings,” and ”To be on the move, traveling, change, and excitement – that’s the kind of life
I like.” The alpha reliability was .82.
Impulsivity was a subscale of the Karolinska Scales of Personality (KSP). The following six
items were answered on a 4-point Likert scale (1 = never agree, 2 = rarely agree, 3 = sometimes
agree, and 4 = always agree) ”I have a tendency to act on the spur of the moment without really
thinking ahead,” ”I usually get so excited over new ideas and suggestions that I forget to check if
there are any disadvantages,” ”I often throw myself too hastily into things,” ”I am a very
conscientious person (rev.),” ”I usually talk before I think,” and ”I consider myself an impulsive
person.” Alpha reliability was .74.
Risk taking. The risk taking scale contained items from different self-report instruments to
measure psychopathic traits. These items have been compiled and used previously (Gustafson &
Ritzer, 1995). Unless otherwise stated, youths responded on a 5-point scale from strongly
disagree (1) to strongly agree (5). The items were as follows: ”I would do almost anything on
dare;” “I dare to do things that others do not” (response scale: 1 = true, 2 = false); ”I enjoy
traveling in cars at high speed;” ”I have often done something dangerous just for the thrill of it;”
”I enjoy taking chances;” ”I would never do something dangerous just for the thrill of it”
(reversed); ”I enjoy gambling for large stakes;” ”I enjoy doing wild things;” and ”I would do
almost anything on a dare” versus “I tend to be a fairly cautious person” (dichotomized measure).
The alpha reliability was .83.
PAPER I
Preexisting Differences as an Explanation for the Interaction between
Youth Center Attendance, Pubertal Timing, and Problem Behavior
16
Female
Pubertal Timing and Problem Behavior
56
Results
Personality characteristics such as thrill seeking, impulsivity, risk taking, and sensation seeking
might be possible explanations why girls commit offences, get involved with boys and overly
involved with peers, or choose to spend their leisure time at the youth centers. These personality
characteristics might be connected with an adventurous life style, and in fact, might explain most
of the interrelations we have observed in this study (and the previous studies) and provide a
reasonable rival hypothesis to the model proposed here.
Table 5.4.
Correlating delinquency, romantic relationships with boys, peer relationships, and youth center
attendance with personality characteristics. The 1998 cohort.
________________________________________________________________________
Delinquency
Romantic
Peer
Center
_____________________________________________________________________
Thrill seeking
.34 ***
.24 ***
.38 ***
.20
Impulsivity
.35 ***
.24 ***
.27 ***
.08
Risk taking
.44 ***
.37 ***
.39 ***
.20
Sensation seeking
.38 ***
.27 ***
.40 ***
.25
________________________________________________________________________
*** p < .001
***
n.s.
***
***
As shown in Table 5.4, there are grounds for believing this. All four personality measures
were quite highly correlated with delinquency. Moreover, all of them were significantly
associated with romantic relationships with boys and with peer involvement. Finally, all of these
measures except impulsivity correlated significantly with youth center attendance.
Table 5.5.
Hierarchical regression analysis predicting delinquency at age 15 from personality
measures associated with an adventurous life style and pubertal timing. The 1998 cohort.
________________________________________________________________________
Variables in the Model
beta
p value
________________________________________________________________________
Model 1.
Main effects of personality measures1
Thrill seeking
Impulsivity
Risk taking
Sensation seeking
.05
.11
.26
.16
n.s.
.040
< .001
.003
Model 2.
Add pubertal timing2
Thrill seeking
.05
n.s.
Impulsivity
.09
.070
Risk taking
.25
< .001
Sensation seeking
.17
.002
Pubertal timing
-.19
< .001
________________________________________________________________________
1
Model F (4,456) = 34.26, p < .001; R2 = .23 2Model F (5,456) = 33.27, p < .001; R2 = .27
17
Female Pubertal Timing and Problem Behavior
57
Table 5.6.
Hierarchical regression predicting delinquent behavior from center attendance, romantic
relationships with boys, pubertal timing, the interactions among these three predictors, and
risk prone personality. The 1998 cohort.
________________________________________________________________________
Variables in the Model
Excluding
Including
risk prone personality2
risk prone personality1
beta
p value
beta
p value
_____________________________________________________________________
Risk prone personality
.40
< .001
Romantic relationships
.11
.026
-.01
.865
Center attendance
.22
<.001
.15
< .001
Pubertal timing
-.18
<.001
-.17
< .001
Romantic x Center
-.05
.369
-.03
.562
Romantic x Puberty
-.14
.008
-.13
.010
Center x Puberty
-.04
.350
-.06
.140
Romantic x Center x Puberty -.15
.009
-.13
.009
________________________________________________________________________
1
Model F (7,487) = 19.60, p < .001; R2 = .22
2
Model F (8,487) = 33.09, p < .001; R2 = .36
In order to examine whether controlling for personality characteristics eliminates the effects
of the three-way interactions, we conducted new analyses of how pubertal timing, center
attendance, and heterosexual (or peer) involvement predicts delinquency, and in these analyses
we entered a broad measure of risk prone personality as a another main predictor (we aggregated
the scales thrill seeking, impulsivity, risk taking, and sensation seeking; the combined measure
had an alpha reliability of .81). The results from the model examining heterosexual involvement
are reported in Table 5.6. When the measure of risk prone personality was entered as a predictor
for delinquency along with pubertal maturation, center attendance, and heterosexual involvement,
risk prone personality was the most important predictor. Despite this, pubertal maturation
remained a significant predictor. The slope of the interaction term Heterosexual involvement x
PAPER I
These personality factors, though, do not seem to rival our basic model about the connection
between pubertal timing and juvenile delinquency. First, two of the four personality measures,
thrill seeking and sensation seeking, were not significantly associated with pubertal timing
whereas the other two correlated with pubertal timing on a low level (r = .11, p < .05 and r = .10,
p < .05 for impulsiveness and risk taking, respectively). Second, in a multiple regression analysis
with delinquency as the dependent variable, when all four personality measures were entered in
the first step, pubertal timing still predicted delinquent behavior. As reported in Table 5.5, the
beta coefficient for pubertal timing was -.19 in this analysis, which included the four personality
measures, as compared with -.24 when pubertal timing was entered as the only predictor. Of the
four personality measures, risk taking was the strongest predictor of delinquent behavior in this
analysis.
It can be argued that even though these personality factors do not explain why pubertal timing
is linked to delinquency, they might explain the interactions between pubertal timing and peer
relationships. Here we are arguing about a predictive impact for a constellation of three factors:
pubertal timing, center attendance, and heterosexual (or peer) involvement.
58
18
Female
Pubertal Timing and Problem Behavior
Center attendance x Puberty was almost identical to what it had been before risk prone
personality was entered.
Table 5.7.
Hierarchical regression predicting delinquent behavior from center attendance, peer involvement,
pubertal timing, the interactions among these three predictors, and risk prone personality. The 1998
cohort.
________________________________________________________________________
Variables in the Model
Excluding
Including
risk prone personality1
risk prone personality2
beta
p value
beta
p value
_____________________________________________________________________
Risk prone personality
.33
< .001
Peer involvements
.30
<.001
.18
< .001
Center attendance
.09
.092
.04
.468
Pubertal timing
-.12
.007
-.10
.010
Peers x Center
.04
.437
.08
.094
Peers x Puberty
-.32
<.001
-.27
< .001
Center x Puberty
.10
.077
.07
.199
Peers x Center x Puberty
-.24
<.001
-.21
< .001
________________________________________________________________________
1
Model F (7,482) = 31.60, p < .001; R2 = .32
2
Model F (8,482) = 39.82, p < .001; R2 = .39
We did a similar analysis for peer interaction (see Table 5.7). Risk prone personality, peer
involvement, center attendance and pubertal timing were entered as predictors of delinquent
behavior, and the interaction terms involving peer involvement, center attendance, and pubertal
timing were also entered. Risk prone personality was the strongest predictor, followed by peer
involvement, and pubertal timing. There were two significant interactions: the two-way
interaction, Peer involvement x Pubertal timing, and the three-way interaction. A plot of these
showed the same findings as reported earlier. In summary, the conclusions drawn in Chapter IV
did not change when risk taking was included as a possible alternative explanation.
Discussion
In view of the evidence that has accumulated over the years that early maturing girls are more
involved in normbreaking behavior than later-maturing girls and more involved in romantic
relationships, the question is whether this can be explained by differences in social adjustment
problems earlier in life. Here we tested the idea that different kinds of problems that exist in late
childhood (behavior, school, and social relations problems) either account for the associations we
find between pubertal timing and mid-adolescent normbreaking and heterosexual relationships, or
that these early problems among early developed girls are accentuated by puberty. The results
show that few of the age 10 problems were related to pubertal timing. Nor were the problems at
age 13 – at a time when about half of all girls were post-pubertal. Early pubertal timing was
significantly associated with teacher reported motor restlessness at age 10 and motor restlessness
and concentration difficulties at age 13, but these differences between the early and later
developing girls did not much reduce the association between pubertal timing and mid-adolescent
normbreaking or between pubertal timing and mid-adolescent heterosexual relationships. Hence,
19
Female Pubertal Timing and Problem Behavior
59
PAPER I
we conclude that differences in mid-adolescent normbreaking and heterosexual relationships
between early and late maturing girls cannot be accounted for by factors that are present already
in late childhood.
A possible limitation of this study is that our measures of problems in late childhood were not
collected at a time when all of the girls were prepubertal. The time from the first signs of puberty
to adult height ranges over many years. One would need data from age 7 or 8 years (before the
present longitudinal study began), in order to make a strict test of whether there are prepubertal
factors associated both with pubertal timing and mid-adolescent problem behavior. On the other
hand, the earlier the collection of such data, the lower would be the magnitude of the associations
between early problem predictors and later problem behavior.
There might have been other conditions earlier than age 10 that might explain the differences
in normbreaking and romantic relationships with boys between early- and later developed girls.
We cannot rule this out. In a recent study, however, early pubertal timing preceded increases in
problem behavior and girls who experienced problems before puberty did not mature earlier than
other girls (Burt et al., 2006). Furthermore, we are quite confident that we have captured relevant
predictive childhood conditions. We examined various types of factors, as reported by the girls,
their parents, peers at school, and their teachers, that we know are linked to delinquency
development from previous studies in the same longitudinal sample (Klinteberg, Anderson,
Magnusson, & Stattin, 1993; Stattin & Magnusson, 1989). We investigated differences between
the early and late maturing girls with respect to social behavior (aggressiveness, hyperactivity,
disharmonious behavior, behaviors that evoke parental concerns like pocket-money, TVwatching, doing homework, curfew time, peer relationships, smoking, alcohol or sniffing, and
disobedience, achievement related information (grades, intelligence, school fatigue, and selfreported school satisfaction), and measures of peer relationships (self-reported satisfaction,
sociometric and self-report information about popularity). Teacher-rated motor restlessness was
the only measure at age 10 that was significantly associated with pubertal timing, and teacherrated motor restlessness and concentration problems were the only variables that were related to
pubertal timing at age 13. In this context it should be noted that previous reviews have provided
mixed findings on the relationship between concentration problems and restlessness, on the one
hand, and, pubertal timing on the other (Buchanan et al., 1992). This study converges on the
conclusion that late childhood adjustment in different respects is not likely to explain the link
between pubertal timing and problem behavior in mid-adolescence. Nor do early adjustment
problems seem to be accentuated by pubertal timing.
60
20
Female
Pubertal Timing and Problem Behavior
Chapter VI: General discussion
Girls who go through puberty earlier than their peers tend to have higher levels of problem
behavior in adolescence. This effect seems to depend, in part, on everyday contexts such as
school, family, and neighborhood (Caspi, 1995; Ge et al. 2002, Ge, Natsuaki, Jin,et al., 2007;
Obeidallah et al., 2004). The idea that early maturing girls associate with older peers and
boyfriends and are drawn into problem behavior by them has often been used as a way to
interpret the findings, but has less often been tested a priori. In this programmatic series of
studies, we drew upon and elaborated previous work on the role of contexts in amplifying the
negative consequences of early pubertal maturation (Caspi, 1995; Ge et al. 2002, Ge, Natsuaki,
Jin,et al., 2007; Obeidallah et al., 2004) and made predictions about girls’ peer relationships and
the links between early maturity and problem behavior that should be found in different contexts.
The results converge to show that under societal and everyday-life conditions that make it easy
for early maturing girls to affiliate with older peers and boyfriends, early maturing girls do tend
to form associations with older peers, and early maturation is linked to problem behavior,
particularly for girls who are most involved with peers and boys. Furthermore, this is especially
true for girls who spend a lot of time in places where delinquent youths tend to congregate.
Notably, girls’ risk-prone characteristics do not provide an alternative explanation. These results
support the peer socialization hypothesis and an extension of the contextual amplification
hypotheses.
This monograph makes several unique contributions to the literature on girls’ early puberty
and problem behavior. One is the extension of the contextual amplification hypothesis and
demonstration of a mechanism through which it might work. In previous descriptions, contextual
amplification was hypothesized to occur in adverse contexts such as impoverished neighborhoods
or harsh family environments (Ge et al., 2002, Ge, Natsuaki, Jin,et al., 2007; Obeidallah et al.,
2004). In our extension of this idea, the contexts that amplify the link between early puberty and
problem behavior might not be defined as adverse from an objective point of view. They need
only have characteristics that facilitate the working of a given mechanism—in this case, peer
socialization. Although we have applied this extension to the peer socialization mechanism, it
might be equally relevant to other explanations.
Another unique contribution of this monograph is to offer a perspective on puberty as a
nonpathological process and the link between early puberty and problem behavior as a
consequence of normal peer processes. In much of the existing literature, puberty is assumed to
be a diathesis or vulnerability. Scholars have applied developmental psychopathology models
such as diathesis-stress models (e.g., Caspi & Moffitt, 1991; Ge, Natsuaki, Jin,et al., 2007). A
diathesis is a pre-existing enduring vulnerability or a predisposition to experience a certain
problem. Stress is something that interferes with people’s homeostasis and is temporary.
According to diathesis-stress models (Walker, Downey, & Bergman, 1989), periods of stress,
such as major life transitions, accentuate prior difficulties and this leads to further problems.
Thus, if a person has a diathesis, or is prone to developing problems, less stress will be needed for
the individual to develop problems. If a person is not prone to developing problems, or lacks a
diathesis, then more stress will be needed for the individual to develop problems. Although the
concept of diathesis-stress is seldom explicitly used, several researchers have used diathesisstress models to explain their findings, conceptualizing puberty either as a diathesis (e.g., Ge,
Natsuaki, Jin,et al., 2007) or a stressor (Caspi & Moffit, 1991). In one model, early puberty is
viewed as a diathesis and an adverse psychosocial context is viewed as a stressor that increases
the risk that early puberty will have a negative influence on behavior (Ge, Natsuaki, Jin,et al.,
21
Female Pubertal Timing and Problem Behavior
61
PAPER I
2007). In the other model, early puberty is viewed as a stressor that increases the likelihood of
problems that existed before puberty being magnified (Caspi & Moffitt, 1991) or that creates
additional stress in the process of adapting to the physical and social changes of puberty
(Kaltiala-Heino, Marttunnen, Rantanen, & Rimpelä, 2003). Either way, however, early puberty is
a liability. In this series of studies, we did not view early puberty as a diathesis. We did not argue
that it is a vulnerability to the development of delinquency if stress arises. Nor do we view early
puberty as a stressor that could magnify earlier problems. Rather, we conceptualized the link
between early puberty and problem behavior as a result of the normal process of developing
romantic relationships in adolescence and the normal process of being attracted to similar others.
This view of early puberty as a nonpathological event that sets into action normal peer processes
is, as far as we know, a new way of thinking about early puberty and its connection to problem
behavior.
What are the advantages of the peer socialization hypothesis compared with the other
hypotheses for explaining the link between early puberty and girls’ problem behavior? The major
advantage is its specific explanations of the mechanisms linking early maturation with problem
behavior. The off-time and early-timing hypotheses explain the link to internalizing problems,
but the mechanisms through which internal problems are transferred into external problems are
not clearly described. According to the accentuation hypothesis, puberty enlarges individual
differences that existed before puberty, but why this accentuation should be especially true for the
early rather than later developing girls has not been spelled out. The contextual amplification
hypothesis posits that the association between pubertal timing and problem behavior should be
strengthened if girls live in adverse contexts, but why early maturing girls should be more
vulnerable to adverse contexts than other girls is not clearly described. The peer socialization
hypothesis rests on two assumptions, which have ample support in the literature. The first is that
early puberty is associated with engaging with older peers and, in particular, engaging with older
boys. The second is that early romantic relationships often are associated with engaging in
normbreaking behaviors. The advantage of the peer socialization hypothesis, then, is that it spells
out the peer selection and socialization processes linking early puberty and problem behavior.
Some limitations of the approach taken in this monograph should be mentioned. First, we
have used one measure of physical maturation—age of menarche—in all the studies. Although
there are good reasons for doing so, as spelled out in Chapter II, we cannot say whether the
findings reported here might be the same with other indicators of pubertal development, such as
breast development, growth spurt, or other indices of the development of primary or secondary
sexual characteristics. These outwardly visible aspects of physical maturation are often assumed
to be a part of the explanation of the implications pubertal timing might have. According to a
social mediation hypothesis (Ge, Natsuaki, Jin,et al., 2007), early puberty evokes certain
behaviors from others, which might lead to adjustment difficulties on behalf of the early maturing
girl. If it is so, then the overt aspects of puberty (e.g., breast development and growth spurt)
might be differently related to the peer socialization processes and adjustment than the covert
aspect (i.e., menarche), used in these studies. Future studies that include several aspects of the
pubertal processes will be able to tease apart these explanations and specify the processes
involved. A further limitation is that we did not include socioeconomic status (SES) in any of the
analyses. In North American samples, pubertal timing is linked to SES, with girls from lower
socioeconomic strata maturing at an earlier age (Ellis & Essex, 2007). Given that SES is linked to
problem behavior, it could be a potential third variable explaining the link between early pubertal
timing and problem behavior. However, pubertal timing does not appear to be associated with
62
22
Female
Pubertal Timing and Problem Behavior
SES in Scandinavian samples (Bratberg, 2007). This minimizes the chances that the findings in
this monograph could be explained by girls’ SES.
The major strength of this group of studies is the theory-driven approach in which predictions
were made about distal (societal) and proximal (school and free-time) contexts and tested using a
number of different cohorts. In the societal comparison, Slovakia was selected in advance for
having characteristics that differed from Sweden in theoretically relevant ways. Similarly, the
Swedish youth center context was chosen for having theoretically relevant characteristics. The
use of different cohorts, different measures, and different types of contexts and finding
convergent results gives us confidence in the main ideas set forth here.
To what extent could processes not considered in these studies provide alternative
explanations for the results? Several explanations have been offered in order to explain the link
between early puberty and problem behavior. One of the most common explanations concerns
internal stress, which is thought to arise from experiencing puberty early and to eventually lead to
adjustment difficulties (Brooks-Gunn et al., 1985; Tschann et al., 1994; Weisner & Ittel, 2002).
Previous research, however, does not support this idea (Tschann et al., 1994; Weisner & Ittel,
2002). Based on these studies, it seems unlikely that internal distress would provide an alternative
explanation for our findings.
Another explanation that some scholars have suggested is that psychosocial processes in the
family bring on early puberty (see Ellis, 2004, for an extensive review), and these might also
explain the link to problem behavior. For instance, conflicts in the family and father absence have
been linked with earlier timing of puberty (Belsky, Steinberg, Houts, Friedman, Dehart,
Cauffman, et al., 2007; Ellis, 2004). The psychosocial acceleration theory says that early stress in
the family leads to early timing of reproductive strategies, including early puberty and also early
onset of sexual behavior. According to the paternal investment theory, which is a related, more
specific, theory, father absence and stepfather presence before puberty contribute to early
pubertal timing and early onset of various types of reproductive behaviors (e.g., early age at first
intercourse and early age at first child). This is viewed as an evolutionary strategy. Thus, early
processes in the family could potentially explain why early maturing girls display more problem
behavior.
Given that certain psychosocial conditions that existed before puberty lead to an earlier
timing of the pubertal process, could the very same conditions act as third variables such that they
would also cause adolescent problem behavior and thus eliminate the link between pubertal
timing and problem behavior? This was the focus of Chapter IV. We were interested in possible
alternative explanations of the link between early puberty and problem behavior. First, we tested
whether potential differences in normbreaking behavior in mid-adolescence between early and
later developed girls could be traced back to individual differences in problem behavior at age 10,
when most of the girls were pre-pubertal, or at least pre-menarcheal. We also investigated the
accentuation hypothesis by testing whether potential early differences between early and late
maturing girls were strengthened as a result of pubertal timing. The peer socialization hypothesis
links girls’ behavior problems to romantic relationships with boys. Therefore, we examined
whether normbreaking behavior and heterosexual involvement could be explained by prepubertal factors. We failed, however, to find support for these two ideas. First, early developing
girls did not seem to have more early problems than the late maturing girls. Second, early puberty
did not seem to accentuate pre-existing problems. It should be noted that menarche is a relatively
late event in the pubertal process and that the early phases of puberty might have already had an
impact on some of the conditions that we refer to as “pre-existing.” To the extent that this is so
and that puberty increases problems, pre-existing problems might have appeared larger than they
23
Female Pubertal Timing and Problem Behavior
63
PAPER I
are in reality. This does not weaken our conclusion that early developing girls did not have more
early problems than late maturing girls, but it could weaken our conclusions about accentuation,
because some of the accentuation might have already occurred. With this caveat, however, we
conclude that early differences between early and late developing girls do not seem to be a strong
alternative explanation of the findings presented in this monograph.
The findings in this monograph raise many questions for future research. What is particularly
urgent is further clarification of the role of context in the link between pubertal timing and
problem behavior and in order for peer socialization to occur. We need to achieve a better
understanding of the social context in which maturing girls spend their time. One of the
remaining unanswered questions is how early maturing girls are brought into older peer
networks. Are the girls or the older peers the driving force behind this process? Are some peer
groups more open than others to younger, early maturing girls, and are there differences between
boys who form relationships with younger, early maturing girls and those who form relationships
with girls their own ages? Another question is how an individual’s maturational timing affects the
peer group. Are girls who mature early marginalized by their peers? Previous research, including
this monograph, has only focused on how the peer group influences the maturing girl. These are
some examples of issues concerning context that should be addressed. Furthermore, more
knowledge is needed about factors associated with resilience for the sake of developing
potentially effective preventive interventions. Understanding why some early maturing girls who
have early romantic relationships and who are in adverse contexts do not develop problem
behavior is a pivotal part of such an endeavor.
How should we prevent early maturing girls from developing problem behavior in
adolescence? This is a difficult question to answer. We think that it is important to provide
accurate information about the psychosocial meaning of early pubertal maturation to
professionals working with adolescent girls and also to their parents. Graber et al. (1997)
reasoned that “it is unlikely that parents are aware of the risks of early maturation for girls” (p.
1774) but more recent research (e.g., O’Sullivan et al., 2000) has shown that parents do believe
that early puberty constitutes a risk in girls’ development. Therefore it is even more important to
be able to provide them with accurate information about the pubertal process and its psychosocial
meaning. Furthermore, professionals and parents should be informed about which circumstances
and under what conditions early puberty constitutes a risk for problem development. Forbidding
early maturing girls to have older peers and boyfriends is not the preventive effort that we would
recommend, because there is no evidence that parents attempts to prohibit specific peer
relationships are successful (see Mounts, 2000). Parents might be able to indirectly influence the
peer contexts their girls choose, however. In a recent study, youths who had good feelings about
the home context, who were treated with respect, and who were able to influence family
decisions were likely to spend their free time in adult-led, structured activities, whereas those
who had negative experiences at home were likely to drop out of structured activities and begin
loitering on the streets (Persson, Kerr, & Stattin, 2007). Research also shows that youths in
structured activities are lower on problem behavior than other youths. Therefore, creating a home
atmosphere in which girls feel valued and respected is one way in which parents might be able to
indirectly influence the type of peers to which girls are exposed. More research is needed,
however, to get knowledge about what can be done to prevent girls from acquiring problem
behaviors when they associate with older peers and boyfriends.
Future models about how pubertal timing affects behavior must recognize the historical,
social and cultural embeddedness of adolescents’ behavior – and they have to be contextual.
Whether or not early- and later-developed girls differ with respect to interpersonal characteristics,
64
24
Female
Pubertal Timing and Problem Behavior
emotional, and behavioral characteristics is likely to depend on and be moderated by contextual
conditions operating at different levels. Developmental psychology is concerned with identifying
the factors that operate behind how people think, feel, act and react across the life span, and
understands the mechanisms by which they operate (Magnusson & Stattin, 2006). Up to the
present time, scientific query about how puberty affects behavior has mainly been concerned with
the first issue, to identify the operating factors. As can be seen in reviews by Susman et al.
(2003), Alsaker (1995b, 1996) and Stattin and Magnusson (1990), empirical studies over almost a
century have given much information about which features differentiate early from late maturing
girls. But we still have limited information about the mechanisms by which they operate. A
number of models have been presented in the literature, such as the off time, early timing,
accentuation, and the peer socialization hypothesis. These ideas have often been largely
acontextual, though, and because of this have not permitted more detailed predictions. Here we
have argued that without knowledge about setting characteristics and the operation of the social
context at different levels, limited progress will be achieved about the role of biological factors in
development. Future models of pubertal timing in girls must be interactive models, and must be
able to make specific predictions of person-situation interactions. If there is anything that this
series of studies shows it is that contexts play the deciding role for early developing girls.
25
Female Pubertal Timing and Problem Behavior
65
References
PAPER I
Abernethy, E. M. (1925). Correlations in physical and mental growth. Journal of Educational
Psychology, 16, 539-546.
Ackard, D. M., & Peterson, C. B. (2001). Association between puberty and disordered eating,
body image, and other psychological variables. International Journal of Eating Disorders,
29, 187-194.
Adler, P. A., & Adler, P. (1988). Intense loyalty in organizations: A case study of college
athletics. Administrative Science Quarterly, 33, 401-417.
Alsaker, F. D. (1995a). Is puberty a critical period for socialization? Journal of Adolescence, 18,
427-444.
Alsaker, F. D. (1995b). Timing of puberty and reactions to pubertal changes. In M. Rutter (Ed.),
Psychosocial disturbances in young people. (pp. 37-82). Cambridge: Cambridge University
Press.
Alsaker, F. D. (1996). Annotation: The impact of puberty. Journal of Child Psychology and
Psychiatry, 37, 249-258.
Andershed, H., Kerr, M., Stattin, H., & Levander, S. (2002). Psychopathic traits in non-referred
youths: A new assessment tool. In E. Blaauw, & L. Sheridan (Eds.), Psychopaths: Current
International Perspectives (pp. 131-158). The Hague: Elsevier.
Aro, H., & Taipale, V. (1987). The impact of timing of puberty on psychosomatic symptoms
among fourteen- to sixteen-year-old Finnish girls. Child Development, 58, 261-268.
Backteman, G., & Magnusson, D. (1978). Longitudinal stability of personality characteristics:
Ratings. Reports from the Department of Psychology, Stockholm University, Sweden,
Report No 537. Hur skrivs denna ref?
Baumeister, L. M., Flores, E., & Marín, B. V. (1995). Sex information given to Latina
adolescents by parents. Health Education Research, 10, 233-239.
Belsky, J., Steinberg, L. D., Houts, R. M., Friedman, S. L., Dehart, G, Cauffman, E., et al. (2007).
Family rearing antecedants of pubertal timing. Child Development, 78, 1302-1321.
Bergsten-Brucefors, A. (1976). A note on the accuracy of recalled age at menarche. Annals of
Human Biology, 3, 71-73.
Biehl, M. C., Natsuaki, M. N., & Ge, X. (2007). The influence of pubertal timing on alcohol use
and heavy drinking trajectories. Journal of Youth and Adolescence, 36, 153-167.
Bingham, C. R., & Crockett, L. J. (1996). Longitudinal adjustment patterns of boys and girls
experiencing early, middle, and late sexual intercourse. Developmental Psychology, 32,
647-658.
Brim, O. G., Jr., & Ryff, C. D. (1980). On the properties of life events. In P. B. Baltes & O. G.
Brim, Jr. (Eds.), Life-span development and behavior (Vol. 3). (pp. 376-388). New York:
Academic Press.
Brooks-Gunn, J. (1987). Pubertal processes and girls’ psychological adaptation. In R. M. Lerner
& T. T. Foch (Eds.), Biological-psychosocial interactions in early adolescence. (pp. 123153). Hillsdale, NJ: Lawrence Erlbaum Associates.
Brooks-Gunn, J., & Graber, J. A. (1994). Puberty as a biological and social event: Implications
for research on pharmacology. Journal of Adolescent Health, 15, 663-671.
Brooks-Gunn, J., Petersen, A. C., & Eichorn, D. (1985). The study of maturational timing effects
in adolescence. Journal of Youth and Adolescence, 14, 149-161.
Brooks-Gunn, J., Warren, M. P., Rosso, J., & Gargiulo, J. (1987) Validity of self-report measures
of girls’ pubertal status. Child Development, 58, 829-841.
66
26
Female
Pubertal Timing and Problem Behavior
Brown, J. D., Halpern, C. T., & L’Engle, K. L. (2005). Mass media as a sexual super peer for
early maturing girls. Journal of Adolescent Health, 36, 420-427.
Buchanan, C. M., Eccles, J. S., & Becker, J. B. (1992). Are adolescents the victims of raging
hormones? Evidence for activational effects of hormon on moods and behavior at
adolescence. Psychological Bulletin, 111, 62-107.
Burt, S., McGue, M., DeMarte, J., Krueger, R. F., & Iacono, W. G. (2006). Timing of menarche
and the origins of conduct disorder. Archives of General Psychiatry, 63, 890-896.
Campbell, A. C. (1980). Friendship as a factor in male and female delinquency. In H. C. Foot, A.
J. Chapman, & J. R. Smith (Eds.), Friendship and social relations in children (pp. 365389). New York: Wiley.
Caspi, A. (1995). Puberty and the gender organization of schools: How biology and social
context shape the adolescent experience. In L. J. Crockett & A. C. Crouter (Eds.), Pathways
through adolescence: Individual development in relation to social contexts. The Penn State
series on child & adolescent development (pp. 57-74). Hillsdale, NJ: Lawrence Erlbaum
Associates.
Caspi, A., Lynam, D., Moffitt, T.E., & Silva, P. (1993). Unraveling girls’ delinquency:
Biological, dispositional, and contextual contributions to adolescent misbehavior.
Developmental Psychology, 29, 19-30.
Caspi, A., & Moffitt, T. E. (1991). Individual differences are accentuated during periods of social
change: The sample case of girls at puberty. Journal of Personality and Social Psychology,
61, 157-168.
Cauffman, E., & Steinberg, L. (1996). Interactive effects of menarcheal status and dating on
dieting and disordered eating among adolescent girls. Developmental Psychology, 32, 631635.
Celio, M., Karnik, N. S., & Steiner, H. (2006). Early maturation as a risk factor for aggression
and delinquency in adolescent girls: a review. International Journal of Clinical Practice,
60, 1254-1262.
Cavanagh, S. E. (2004). The Sexual Debut of Girls in Early Adolescence: The Intersection of
Race, Pubertal Timing, and Friendship Group Dynamics. Journal of Research on
Adolescence, 14, 285-312.
Chilman, C. S. (1986). Some psychosocial aspects of adolescent sexual and contraceptive
behaviors in a changing American society. In J. B. Lancaster, & B. A. Hamburg (Eds.),
School-age pregnancy and parenthood: Biosocial dimensions (pp. 191-217). New York:
Aldine De Gruyter.
Clausen, J. A. (1975). The social meaning of differential physical and sexual maturation. In S. E.
Dragastin amd G. H. Elder (Eds.), Adolescence in the life cycle: Psychological change and
social context (pp. 25-47). Washington, DC: Hemisphere.
Coleman, J. S. (1961). The adolescent society. Oxford, England: Free Press of Glencoe.
Coleman, L., & Coleman, J. (2002). The measurement of puberty: A review. Journal of
Adolescence, 25, 535-550.
Comings, D. E., Muhleman, D., Johnson, J. P. & MacMurray, J. P. (2002). Parent-daughter
transmission of the androgen receptor gene as an explanation of the effect of father absence
on age of menarche. Child Development, 73, 1046-1051.
Comite, F., Pescovitz, O. H., Sonis, W. A., Hench, K., McNemar, A., Klein, R., P., Loriaux, D.
L., & Cutler, G. B. (1987). Premature adolescence: Neuroendocrine and psychosocial
studies. In R. M. Lerner & T. T. Foch (Eds.), Biological-psychosocial interactions in early
adolescence. (pp. 155-172). Hillsdale, NJ: Lawrence Erlbaum Associates.
27
Female Pubertal Timing and Problem Behavior
67
PAPER I
Compian, L., Gowen, L. K., & Hayward, C. (2004). Peripubertal girls’ romantic and platonic
involvement with boys: Associations with body image and depression symptoms. Journal
of Research on Adolescence, 14, 23-47.
Conger, R. D., & Rueter, M. A. (1996). Siblings, parents, and peers: A longitudinal study of
social influences in adolescent risk for alcohol use and abuse. In G. Brody, (Ed.), Sibling
relationships: Their causes and consequences. (pp.1-30). Westport, CT, US: Ablex
Publishing.
Connolly, S. D., Paikoff, R., & Buchanan, C. M. (1996). Puberty: The interplay of biological and
psychosocial processes in adolescence. In G. R. Adams & R. Montemayor (Eds.),
Psychosocial development during adolescence (Vol. 8, pp. 259-299). Thousand Oaks, CA:
Sage.
Cooper, M. L., Wood, P. K., Orcutt, H. K., & Albino, A. (2003). Personality and the
predisposition to engage in risky or problem behaviors during adolescence. Journal of
Personality and Social Psychology, 84, 390-410.
Costa, F. M., Jessor, R., Donovan, J. E., & Fortenberry, J. D. (1995). Early initiation of sexual
intercourse: The influence of psychosocial unconventionality. Journal of Research on
Adolescence, 5, 93-121.
Costello, E. J., Sung, M., Worthman, C., & Angold, A. (2007). Pubertal maturation and the
development of alcohol use and abuse. Drug and Alcohol Dependence, 88, s50-59.
Crafoord, K. (1986). Flickproblem och problemflickor i tidiga tonåren. [Problem behavior and
problem-girls in early youth]. Individuell Utveckling och Miljö, Stockholms Universitet,
Rapport Nr 65.
Crockett, L. J., Bingham, C. R., Chopak, J. S., & Vicary, J. R. (1996). Timing of first sexual
intercourse: The role of social control, social learning, and problem behavior. Journal of
Youth and Adolescence, 25, 89-111.
Crockett, L., & Petersen, A. C. (1987). Pubertal status and psychosocial development: Findings
from the Early Adolescence Study. In R. M. Lerner & T. T. Foch (Eds.), Biologicalpsychosical interactions in early adolescence: A life-span perspective. (pp. 173-188).
Hillsdale, NJ: Lawrence Erlbaum Associates.
Darroch, J. E., Frost, J. J., & Singh, S. (2001). Differences in teenage pregnancy rates among five
developed countries: The roles of sexual activity and contraceptive use, Family Planning
Perspectives, 33, 244-250.
Dick, D. M., Rose, R. J., Pulkkinen, L. & Kaprio, J. (2001). Measuring puberty and
understanding its impact: A longitudinal study of adolescent twins. Journal of Youth and
Adolescence, 30, 385-400.
Dick, D. M., Rose, R. J., Viken, R. J., & Kaprio, J. (2000). Pubertal timing and substance use:
Associations between and within families across late adolescence. Developmental
Psychology, 36, 180-189.
Dishion, T. J., Patterson, G. R., & Griesler, P. C. (1994). Peer adaptations in the development of
antisocial behavior. In L. R. Huesmann (Ed.), Current perspectives on aggressive behavior
(pp. 61-95). New York: Plenum Press.
Dorn, L. D., Crockett, L. J., & Petersen, A. C. (1988). The relations of pubertal status to
intrapersonal changes in young adolescents. Journal of Early Adolescence, 8, 405-419.
Dornbusch, S. M., Carlsmith, J. M., Gross, R .T., Martin, J. A., Jennings, D., Rosenberg, A., &
Duke, P. M. (1981). Sexual development, age, and dating: A comparison of biological and
social influences upon one set of behaviors. Child Development, 52, 179-185.
Douglas, J. W. B. (1964). The home and the school. London: Macgibbon & Kee.
68
28
Female
Pubertal Timing and Problem Behavior
Douglas, J. W. B., & Ross, J. M. (1964). Age of puberty related to educational ability, attainment
and school leaving age. Journal of Child Psychology and Psychiatry, 5, 185-195.
Dorn, L. D., Dahl, R. E., Woodward, H. R., & Biro, F. (2006). Defining the boundaries of early
adolescence: A user's guide to assessing pubertal status and pubertal timing in research with
adolescents. Applied Developmental Science, 10, 30-56.
Dubas, J. S. (2002). Long-term Versus Temporary Effects of Pubertal Development on
Internalizing and Externalizing Problems. Paper presented at the European Association of
Research on Adolescence conference, Oxford, September 2002.
Duke, P. M., Carlsmith, J. M., Jennings, D., Martin, J. A., Dornbusch, S. M., Gross, R. T., &
Siegel-Gorelick, B. (1982). Educational correlates of early and late sexual maturation in
adolescence. Journal of Pediatrics, 100, 633-637.
Duncan, P. D., Ritter, P. L., Dornbusch, S. M., Gross, R. T., & Carlsmith, J. M. (1985). The
effets of pubertal timing on body image, school behavior, and deviance. Journal of Youth
and Adolescence, 14, 227-235.
Edgardh, K. (2002). Sexual behaviour in a low-income high school setting in Stockholm. Internal
Journal of STD & AIDS, 13, 160-167.
Edgardh, K., Lewin, B., & Nilsson, B. R. (1999). Sexual experience and behaviour as reported by
17-year-old girls and boys in Sweden. Scandinavian Journal of Sexology, 2, 41-60.
Ehrhardt, A. A., Meyer-Bahlburg, H. F. L., Bell, J. R., Cohen, S. F., Healey, J. M., Stiel, R.,
Feldman, J. F., Morishima, A., & New, M. I. (1984). Idiopathic precocious puberty in girls:
Psychiatric follow-up in adolescence. Journal of the American Academy of Child
Psychiatry, 23, 23-33.
Eichorn, D. H. (1975). Asynchronizations in adolescent development. In S. E. Dragastin & G. H.
Elder (Eds.), Adolescence in the life cycle: Psychological change and social context. (xi,
324 pp.) Oxford, England: Hemisphere.
Elliott, D. S., & Morse, B. J. (1989). Delinquency and drug use as risk factors in teenage sexual
activity. Youth and Society, 21, 32-60.
Ellis, B. J. (2004). Timing of pubertal maturation in girls: an integrated life history approach.
Psychological Bulletin, 130, 920-958.
Faust, M. S. (1983). Alternative constructions of adolescent growth. In J Brooks-Gunn & A. C.
Petersen (Eds.), Girls at puberty. Biological and psychosocial perspectives. (pp. 105-125).
New York: Plenum.
Fergusson, D. M., Vitaro, F., Wanner, B., & Brendgen, M. (2007). Protective and compensatory
factors mitigating the influence of deviant friends on delinquent behaviours during early
adolescence. Journal of Adolescence, 30, 33-50.
Fidler, J. A., West, R., Jarvis, M. J., & Wardle, J. (2006). Early dating predicts smoking during
adolescence: A prospective study. Addiction, 101, 1805–1813.
Flannery, D. J., Rowe, D. C., & Gulley, B. L. (1993). Impact of pubertal status, timing, and age
on adolescent sexual experience and delinquency. Journal of Adolescent Research, 8, 2140.
Forrest, J. D. (1990). Cultural influences on adolescent reproductive behavior. In J. Bancroft & J.
M. Reinisch (Eds.), Adolescence and puberty (pp 234-253), Oxford: Oxford University
Press.
Friedman, H. L. (1992). Changing patterns of adolescent sexual behavior: Consequences for
health and development. Journal of Adolescent Health, 13, 345-350.
Frisk, M. (1968). Tonårsproblem. En studie av läroverksungdom. (Teenage problems).
Helsingfors: Samfundet Folkhälsan. (Dissertation, not in English).
29
Female Pubertal Timing and Problem Behavior
69
PAPER I
Frisk, M., Tenhunen, T., Widholm, O., & Hortling, M. (1966). Psychological problems in
adolescents showing advanced or delayed physical maturation. Adolescence, 1, 126-135.
Gagnon, J. H. (1983). Age at menarche and sexual conduct in adolescence and young adulthood.
In S. Golub (Ed.), Menarche. The transition from girl to woman. Lexington, MA:
Lexington Books, D. C. Heath.
Garwood, S. G., & Allen, L. (1979). Self-concept and identified problem differences between
pre- and postmenarcheal adolescents. Journal of Clinical Psychology, 35, 528-237.
Ge, X., Brody, G. H., Conger, R. D., Simons, R. L. (2006). Pubertal transition and African
American children’s internalizing and externalizing symptoms. Journal of Youth and
Adolescence, 35, 531-540.
Ge, X., Brody, G. H., Conger, R. D., Simons, R. L., & Murry, V. M. (2002). Contextual
amplification of pubertal transition effects on deviant peer affiliation and externalizing
behavior among African American children. Developmental Psychology, 38, 42-54.
Ge, X., Conger, R. D., & Elder, G. H. (1996). Coming of age too early: Pubertal influences on
girls’ vulnerability to psychological distress. Child Development, 67, 3386-3400.
Ge, X., Conger, R. D., & Elder, G. H. (2001). Pubertal transition, stressful life events, and the
emergence of gender differences in adolescent depressive symptoms. Developmental
Psychology, 37, 404-417.
Ge, X., Elder, G. H., Regnerus, M., & Cox, C. (2001). Pubertal transitions, perceptions of being
overweight, and adolescents’ psychological maladjustment: Gender and ethnic differences.
Social Psychology Quarterly, 64, 363-375.
Ge, X., Jin, R., Natsuaki M.N., Gibbons, F.X., Brody, G.H., Cutrona, C.E., & Simons, R.L.
(2006). Pubertal Maturation and Early Substance Use Risks in African American Children.
Psychology of Addictive Behaviors, 20, 404-414.
Ge, X., Natsuaki, M. N., Jin, R., & Biehl, M. (2007). A Contextual Amplification Hypothesis:
Pubertal Timing and Girls’ Problem Behaviors. Manuscript.
Ge, X., Natsuaki, M. N., Neiderhiser, J. M., & Reiss, D. (2007). Genetic and environmental
influences on pubertal timing: Results from two national sibling studies. Journal of
Research on Adolescence, 17, 767-788.
George, T. P., & Hartmann, D. P. (1996). Friendship networks of unpopular, average, and
popular children. Child Development, 67, 2301-2316.
Gilger, J. W., Geary, D. C., & Eisele, L. M. (1991). Reliability and validity of retrospective selfreports of age pubertal onset using twin, sibling, and college student data. Adolescence, 26,
41-53.
Gibbs, J. T. (1986). Psychosocial correlates of sexual attitudes and behaviors in urban early
adolescent females: Implications for intervention. Journal of Social Work & Human
Sexuality, 5, 81-97.
Giordano, P. C. (1978). Girls, guys, and gangs: The changing social context of female
delinquency. American Journal of Sociology, 91, 1170-1203.
Giordano, P. C., Cernkovich, S. A., & Pugh, M. D. (1986). Friendships and delinquency.
American Journal of Sociology, 91, 1170-1202.
Gold, M., & Petronio, R. J. (1980). Delinquent behavior in adolescence. In J. Adelson (Ed.).
Handbook of adolescent psychology (pp. 495-535). New York: Wiley.
Goldman, R. J., & Goldman, J. D. G. (1982). Children’s sexual thinking. London: Routledge &
Kegan Paul.
70
30
Female
Pubertal Timing and Problem Behavior
Goodson, P., Evans, A., & Edmundson, E. (1997). Female adolescents and onset of sexual
intercourse: A theory-based review of research from 1984 to 1994. Journal of Adolescent
Health, 21, 147-156.
Graber, J. S. (2003). Puberty in context. In C. Hayward (ed.), Gender differences at puberty.
(pp.307-325). New York, NY, US: Cambridge University Press.
Graber, J. A., & Brooks-Gunn, J. (1996). Transitions and turning points: navigating the passage
form childhood through adolescence. Developmental Psychology, 32, 768-776.
Graber, J. A., & Brooks-Gunn, J. (2002). Adolescent girls’ sexual development. In G. M.
Wingood & R. J. DiClemente (Eds.), Handbook of women’s sexual and reproductive
health. Issues in women health. (pp. 21-42). New York, NY: Kluwer Academic/Plenum
Publishers.
Graber, J. A., Brooks-Gunn, J., & Warren, M. (1995). The antecedents of menarcheal age:
Heredity, family environment, and stressful life events. Child Development, 66, 346-359.
Graber, J. A., Brooks-Gunn, J. & Warren, M. P. (2006). Pubertal effects on adjustment in girls:
moving from demonstrating effects to identifying pathways. Journal of Youth and
Adolescence, 35, 413-423.
Graber, J. A., Lewinsohn, P. M., Seeley, J. R., & Brooks-Gunn, J. (1997). Is psychopathology
associated with the timing of pubertal development? Journal of the American Academy of
Child and Adolescent Psychiatry, 36, 1768-1776.
Graber, J. A., Seeley, J. R., Brooks-Gunn, J., & Lewinsohn, P. M. (2004). Is Pubertal Timing
Associated With Psychopathology in Young Adulthood? Journal of the American Academy
of Child and Adolescent Psychiatry, 43, 718-726.
Graber, J. A., Petersen, A. C., & Brooks-Gunn, J. (1996). Pubertal processes: Methods, measures,
and models. In J. A.Graber, J. Brooks-Gunn, & A. C. Petersen (Eds.), Transitions through
adolescence (pp. 23-53), New Jersey: Lawrence Erlbaum Associates.
Gustafson, S. B., & Magnusson, D. (1991). Female life careers: A pattern approach. Hillsdale:
Erlbaum.
Gustafson, S. B., & Ritzer, D. R. (1995). The dark side of normal: A psychopathy-linked pattern
called aberrant self-promotion. European Journal of Personality, 9, 147-183.
Gustafson, S. B., Stattin, H., & Magnusson, D. (1992). Aspects of the development of a career
versus homemaking orientation among females: The longitudinal influence of educational
motivation and peers. Journal of Research on Adolescence, 2, 241-259.
Halpern, C. T., Kaestle, C. E., Hallfors, D. D. (2007). Perceived physical maturity, age of
romantic partner, and adolescent risk behavior. Prevention Science, 8, 1-10.
Halpern, C. T., Udry, J. R., Campbell, B., & Suchindran, C. (1999). Effects of body fat on weight
concerns, dating, and sexual activity: A longitudinal analysis of black and white
adolescent females. Developmental Psychology, 35, 721-736.
Hartup, W. W. (1996). The company they keep: Friendships and their developmental
significance. Child Development, 67, 1-13.
Haynie, D. L. (2003). Contexts of risk? Explaining the link between girls’ pubertal development
and their delinquency development. Social Forces, 82, 355-397.
Iannotti, R. J., Bush, P. J., & Weinfurt, K. P. (1996). Perceptions of friends' use of alcohol,
cigarettes, and marijuana among urban schoolchildren: A longitudinal analysis. Addictive
Behaviors, 21, 615-632.
Jessor, R. (1992). Risk behavior in adolescence: A psychosocial framework for understanding
and action. Developmental Review, 12, 374-390.
31
Female Pubertal Timing and Problem Behavior
71
PAPER I
Jessor, S. L., & Jessor, R. (1975). Transition from virginity to nonvirginity among youth: A
social-psychological study over time. Developmental Psychology, 11, 473-484.
Jessor, S. L., & Jessor, R. (1977). Problem behavior and psychological development: A
longitudinal study of youth. New York: Academic Press.
Johansson, T., & Ritzén, E. M. (2005). Very long term follow-up of girls with early and late
menarche. Endocrine Development, 8, 126-136.
Johnson, B. M., & Collins, W. A. (1988). Perceived maturity as a function of appearance cues in
early adolescence: Ratings by unacquainted adults, parents and teachers. Journal of Early
Adolescence, 8, 357-372.
Jorm, A. F., Christensen, H., Rodgers, B., Jacomb, P. A. & Easteal, S. (2004). Association of
adverse childhood experiences, age of menarche and adult reproductive behavior: Does the
androgen receptor gene play a role? American Journal of Medical Genetics
(Neuropsychiatric Genetics), 125B, 105-111.
Kandel, D. B. (1978). Similarity in real-life adolescent friendship pairs. Journal of Personality
and Social Psychology, 36, 306-312.
Kandel, D. B. (1985). On processes of peer influences in adolescent drug use: A developmental
perspective. Advances in Alcohol and Substance Abuse, 4, 139-163.
Kandel, D. B. (1986). Processes of peer influences in adolescence. In R. K. Silbereisen, K.
Eyferth, & G. Rudinger (Eds.), Development as action in context: Problem behaviors and
normal youth development (pp. 203-227). Heidelberg: Springer-Verlag.
Kaprio, J., Rimpelä, A., Winter, T., Viken, R. J., Rimpelä, M., & Rose, R. J. (1995). Common
genetic influences on BMI and age at menarche. Human Biology, 67, 739-753.
Keel, P. K., Fulkerson, J. A., & Leon, G. R. (1997). Disordered eating precursors in pre- and
early adolescent girls and boys. Journal of Youth and Adolescence, 26, 203-216.
Kerr, M., & Stattin, H. (2000). What parents know, how they know it, and several forms of
adolescent adjustment: Further support for a reinterpretation of monitoring. Developmental
Psychology, 36, 366-380.
Kerr, M., Stattin, H., & Kiesner, J. (2007). Peers and problem behavior: Have we missed
something? In R. C. M. E. Engels, M. Kerr, & H. Stattin (Ed.) Friends, Lovers, and
Groups: Key Relationships in Adolescence (pp. 125 – 153) London: Wiley
Ketterlinus, R. D., & Lamb, M. E. (Eds.) (1994). Adolescent problem behaviors: Issues and
research. Hillsdale, NJ: Erlbaum.
Ketterlinus, R. D., Lamb, M. E., Nitz, K., & Elster, A. B. (1992). Adolescent nonsexual and sexrelated problem behaviors. Journal of Adolescent Research, 7, 431-456.
Klinteberg, B., Anderson, T., Magnusson, D., and Stattin, H. (1993). Hyperactive behavior in
childhood as related to subsequent alcohol problems and violent offending: A longitudinal
study of male subjects. Personality and Individual Differences, 15, 381-388.
Lam, T. H., Shi, H. J., Ho, L. M., Stewart, S. M., & Fan, S. (2002). Timing of pubertal
maturation and heterosexual behavior among Hong Kong Chinese adolescents. Archives of
Sexual Behavior, 31, 359-366.
Lanza, S. T., & Collins, L. M. (2002). Pubertal timing and the onset of substance use in females
during early adolescence. Prevention Science, 3, 69-82.
Leas, L., & Mellor, D. (2000). Prediction of delinquency: The role of depression, risk-taking, and
parental attachment. Behaviour Change, 17, 155-166.
72
32
Female
Pubertal Timing and Problem Behavior
Lenerz, K., Kucher, J. S., East, P. L., Lerner, J. V., & Lerner, R. M. (1987). Early adolescents’
physical organismic characteristics and psychosocial functioning: Findings from the
Pennsylvania Early Adolescent Transitions Study (PEATS). In R. M. Lerner & T. T. Foch
(Eds.), Biological-psychosocial interactions in early adolescence. (pp. 225-247). Hillsdale,
NJ: Lawrence Erlbaum Associates.
Lewin, B. (1982). The adolescent boy and girl: First and other early experiences with intercourse
from a representative sample of Swedish school adolescents. Archives of Sexual Behavior,
11, 417-428.
Lindberg, S. M., Grabe, S., & Hyde, J. S. (2007). Gender, pubertal development, and peer sexual
harassment predict objectified body consciousness in early adolescence. Journal of
Research on Adolescence, 17, 723-742.
Lynne, S., Graber, J. A., Nichols, T., Brooks-Gunn J, Botvin G. (2007). Links between Pubertal
Timing, Peer Influences, and Externalizing Behaviors among Urban Students Followed
Through Middle School. Journal of Adolescent Health, 40, 181.e7–181.e13.
Magnusson, D. (1988). Individual development from an interactional perspective. Hillsdale, NJ:
Erlbaum.
Magnusson, D., Dunér, A., & Zetterblom, G. (1975). Adjustment: A longitudinal study. New
York: Wiley.
Magnusson, D., & Stattin, H. (1998). Person-context interaction theories. In W. Damon & R. M.
Lerner (Eds.), Handbook of child psychology. Volume 1: Theoretical models of human
development (pp. 685-759). New York: Wiley.
Magnusson, D., & Stattin, H. (2006). The person in the environment: Towards a general model
for scientific inquiry. In R. M. Lerner (Ed.), Theoretical models of human development.
Volume 1 of Handbook of Child Psychology (6th ed.). (pp. 400-464). New York: Wiley.
Magnusson, D., Stattin, H., & Allen, V. (1985). Biological maturation and social development: A
longitudinal study of some adjustment processes from mid-adolescence to adulthood.
Journal of Youth and Adolescence, 14, 267-283.
Magnusson, D., Stattin, H., & Allen, V. (1986). Differential maturation among girls and its
relation to social adjustment: A longitudinal perspective. In P. Baltes, D. Featherman, & R.
Lerner (Eds.), Life span development (Vol. 7). New York: Academic Press, pp. 134-172.
Mahoney, J. L. & Stattin, H. (2000). Leisure time activities and adolescent anti-social behavior:
The role of structure and social context. Journal of Adolescence, 23, 113-127.
Mahoney, J. L., Stattin, H., & Magnusson, D. (2001). Youth leisure activity participation and
individual adjustment: The Swedish youth recreation center. International Journal of
Behavioral Development, 509-520.
Martin, C. A., Kelly, T. H., Rayens, M. K., Brogli, B. R., Brenzel, A., Smith, W. J., & Omar, H.
A. (2002). Sensation seeking, puberty, and nicotine, alcohol, and marijuana use in
adolescence. Journal of the American Academy of Child and Adolescent Psychiatry, 41,
1495-1502.
McCabe, M. P. (1984). Toward a theory of adolescent dating. Adolescence, 19, 159-170.
McLaren, L., Hardy, R., & Kuh, D. (2003). Women’s body satisfaction at midlife and lifetime
body size: A prospective study. Health Psychology, 22, 370-377.
McMaster, L., Connolly, J., Pepler, D. J., & Craig, W. M. (2002). Peer to peer sexual harassment
among early adolescents. Development and Psychopathology, 14, 91-105.
Mendle, J., Turkheimer, E., & Emery, R. E. (2007). Detrimental psychological outcomes
associated with early pubertal timing in adolescent girls. Developmental Review, 27, 151171.
33
Female Pubertal Timing and Problem Behavior
73
PAPER I
Metzler, C. W., Noell, J., & Biglan, A. (1992). The validation of a construct of high-risk sexual
behavior in heterosexual adolescents. Journal of Adolescent Research, 7, 233-249.
Meyer-Bahlburg, H. F. L., Ehrhardt, A. A., Bell, J. R., Cohen, S. F., Healey, J. M., Feldman, J.
F., Morishima, A., Baker, S. W., & New, M. I. (1985). Idiopathic precocious puberty in
girls: Psychosexual development. Journal of Youth and Adolescence, 14, 339-353.
Mezzich, A. C., Giancola, P. R., Lu, S. Y., Parks, S. M., Ratica, G. M., & Dunn, M. (1999).
Adolescent females with a substance use disorder: Affiliations with adult male sexual
partners. American Journal on the Addictions, 8, 190-200.
Mezzich, A. C., Tarter, R. E., Hsieh. Y-C., & Fuhrman, A. (1992). Substance abuse severity in
female adolescents. Association between age and menarche and chronological age.
American Journal on Addictions, 1, 217-221.
Moffitt, T. E. (1993). Adolescence–limited and life-course-persistent antisocial behavior. A
developmental taxonomy. Psychological Review, 100, 674-701.
Moffitt, T. E., Caspi, A., Belsky, J., & Silva, P. A. (1992). Childhood experience and the onset of
menarche: A test of a sociobiological model. Child Development, 63, 47-58.
Mounts, N. S. (2000). Parental management of adolescent peer relationships: What are its effects
on friend selection? To appear in K.A. Kerns, J.M. Contreras, & A.M. Neal-Barnett
(Eds.), Family and peers: Linking two social worlds. Westport, CT: Greenwood/Praeger.
Muris P., Meesters C., van de Blom W., & Mayer B. (2005). Biological, psychological, and
sociocultural correlates of body change strategies and eating problems in adolescent boys
and girls. Eating Behaviors, 6, 11–22.
Neugarten, B. L. (1969). Continuities and discontinuities of psychological issues into adult life.
Human Development, 12, 121-130.
Neugarten, B. L., Moore, J. W., & Lowe, J. C. (1965). Age norms, age constraints, and adult
socialization. American Journal of Sociology, 70, 710-17.
Obeidallah, D., Brennan, R. T., Brooks-Gunn, J., & Earls, F. (2004). Links between pubertal
timing and neighborhood contexts: Implications for girl’s violent behavior. Journal of the
American Academy of Child and Adolescent Psychiatry, 43, 1460-1468.
Oliver, J., Bakker, A. B., Demerouti, E., & de Jong, R. D. (2005). Projection of own on others’
job characteristics: evidence for the false consensus effect in job characteristics
informations. International Journal of Selection and Assessment, 13, 63-74.
O’Sullivan, L. F., Heino, F. L., Meyer-Bahlburg, H. F. L., & Watkins, B. X. (2000). Social
cognitions associated with pubertal development in a sample of urban, low-income,
African-American and Latina girls and mothers. Journal of Adolescent Health, 27, 227235.
Otor, S. C. J., & Pandey, A. (1999). Adolescent transition to coitus and premarital childbearing in
Suda: A biosocial context. Journal of Biosocial Science, 31, 361-374.
Patterson, G. R., & Stoolmiller, M. (1991). Replications of a dual failure model for boys’
depressed mood. Journal of Consulting and Clinical Psychology, 59, 491-498.
Patton, G. C., McMorris, B. J., Toumbourou, J. W., Hemphill, S. A., Donath, S., & Catalano, R.
F. (2004). Puberty and the Onset of Substance Use and Abuse. Pediatrics, 114, 300-306.
Persson, E., & Jarlbro, G. (1992). Sexual behavior among youth clinic visitors in Sweden:
knowledge and experiences in an HIV perspective. Genitourinary Medicine, 68, 26–31.
Persson, A., Kerr, M., & Stattin, H. (2004). Why a leisure context is linked to normbreaking for
some girls and not others: Personality characteristics and parent-child relations as
explanations. Journal of Adolescence, 27, 583 - 598.
74
34
Female
Pubertal Timing and Problem Behavior
Peskin, H. (1973). Influence of the developmental schedule of puberty on learning and ego
functioning. Journal of Youth and Adolescence, 2, 273-290.
Petersen, A. C., Leffert, N., & Graham, B. L. (1995). Adolescent development and the emergence
of sexuality. Suicide and Life-Threatening Behavior, 25, 4-17.
Petersen, A. C., & Taylor, B. (1980). The biological approach to adolescence: Biological change
and psychological adaptation. In J. Adelson (Ed.), Handbook of adolescent psychology,
(pp.117-155). New York: Wiley.
Phinney, V. G., Jensen, L. C., Olsen, J. A., & Cundick, B. (1990). The relationship between early
development and psychosexual behaviors in adolescent females. Adolescence, 25, 321-332.
Pickles, A., Pickering, K., Simonoff, E., Silberg, J. Meyer, J, & Maes, H. (1998). Genetic
“clocks” and “soft” events: A twin model for pubertal development and other recalled
sequences of developmental milestones, transitions, or ages at onset. Behavior Genetics, 28,
243-253.
Poppleton, P. K. (1968). Puberty, family size and the educational progress of girls. British
Journal of Educational Psychology, 38, 286-292.
Richards, M. H., & Larson, R. (1993). Pubertal development and emotional well-being among
young adolescent girls and boys. Journal of Research on Adolescence, 3, 145-169.
Rodriguez-Tomé, H., Bariaud, F., Cohen Zardi, M. F., Delmas, C., Jeanvoine, B., & Slylagyi, P.
(1993). The effects of pubertal changes on body image and relations with peers of the
opposite sex in adolescence. Journal of Adolescence, 16, 421-438.
Ross, L., Greene, D., & House, P. (1977). The false consensus effect: An egocentric bias in social
perception and attribution processes. Journal of Experimental Social Psychology, 13, 279301.
Rosenbaum, E., & Kandel, D. B. (1990). Early onset of adolescent sexual behavior and drug
involvement. Journal of Marriage and the Family, 52, 783-798.
Rowe, D. C., & Rodgers, J. L. (1994). A social contagion model of adolescent sexual behavior:
Explaining race differences. Social Biology, 41, 1-18.
Rowe, D. C., Rodgers, J. L., & Meseck-Bushey, S. (1989). An "epidemic" model of sexual
intercourse prevalences for Black and White adolescents. Social Biology, 36, 127-145.
Sampson, R. J., & Laub, J. H. (2003). Life-course desisters? Trajectories of crime among
delinquent boys followed to age 70. Criminology, 41, 555-593.
Schinke, S. P., Botvin, G. J., & Orlandi, M. A. (1991). Substance abuse in children and
adolescents: Evaluation and intervention. Newbury Park, CA: Sage Publications.
Schor, N. (1993). Abortion and adolescence: Relation between the menarche and sexual activity.
International Journal of Adolescent Medicine and Health, 6, 225-240.
Silbereisen, R. K., & Kracke, B. (1993). Variation in maturational timing and adjustment in
adolescence. In S. Jackson & H. Rodriquez-Tomé (Eds.), The social worlds of adolescence
(pp. 67-94). East Sussex, UK: Erlbaum.
Silbereisen, R. K., & Kracke, B. (1997). Self-reported maturational timing and adaptation in
adolescence. In J. Schulenberg and J. L. Maggs (Eds.), Health risks and developmental
transitions during adolescence (pp. 85-109). Cambridge, UK: Cambridge University Press.
Silbereisen, R. K., Petersen, A. C., Albrecht, H. T., & Kracke, B. (1989). Maturational timing and
the development of problem behavior: Longitudinal studies in adolescence. Journal of
Early Adolescence, 9, 247-268.
Siegel, J. M., Yancey, A. K., Aneshensel, C. S., & Schuler, R. (1999). Body image, perceived
pubertal timing, and adolescent mental health. Journal of Adolescent Health, 25, 155-165.
35
Female Pubertal Timing and Problem Behavior
75
PAPER I
Simmons, R. G., & Blyth, D. A. (1987). Moving into adolescence: The impact of pubertal change
and school context. New York: De Gruyter.
Simmons, R. G., Blyth, D. A., & McKinney, K. L. (1983). The social and psychological effects
of puberty on white females. In J. Brooks-Gunn & A. C. Petersen (Eds.), Girls at puberty.
Biological and psychosocial perspectives. New York: Plenum Press.
Simmons, R. G., Blyth, D. A., Van Cleave, E. F., & Bush, D. E. (1979). Entry into early
adolescence: The impact of school structure, puberty, and early dating on self-esteem.
American Sociological Review, 44, 948-967.
Simmons, R. G., Carlton-Ford, S. L., & Blyth, D. A. (1987). Predicting how a child will cope
with the transition to junior high school. In R. M. Lerner, & T. T. Foch (Eds.), Biologicalpsychological interaction in early adolescence. Hillsdale, NJ: Lawrence Erlbaum
Associates.
Simon, A. E., Wardle, J., Jarvis, M. J., Steggles, N., & Cartwright, M. (2003) Examining the
relationship between pubertal stage, adolescent health behaviours and stress. Psychological
Medicine, 33, 1369-1379.
Smith, E. A., Udry, J. R. & Morris, N. M. (1985). Pubertal development and friends: A biosocial
explanation of adolescent sexual behavior. Journal of Health and Social Behavior, 26, 183192.
Sonis, W., Comite, F., Blue, J., Pescovitz, O. H., Rahn, C. W., Hench, K. D., Cutler, G., Loriaux,
D. L., & Klein, R. P. (1985). Behavior problems and social competence in girls with true
precocious puberty. Journal of Pediatrics, 106, 156-160.
Stattin, H., & Kerr, M. (2000). Parental monitoring: A reinterpretation. Child Development, 71,
1072-1085.
Stattin, H., Kerr, M., Mahoney, J., Persson, A. & Magnusson, D. (2005). Explaining why a
leisure context is bad for some girls and not for others. In J. L. Mahoney, R. W. Larson, &
J. S. Eccles (Eds.), Organized activities as contexts of development: Extracurricular
activities, after-school and community programs. (pp. 211-234). Mahwah, NJ: Erlbaum.
Stattin, H., & Magnusson, D. (1989). Social transition in adolescence: A biosocial perspective. In
A. de Ribaupierre (Ed.), Transition mechanisms in child development: The longitudinal
perspective. Cambridge: Cambridge University Press, pp. 147-190.
Stattin, H., & Magnusson, D. (1989b). The role of early aggressive behavior for the frequency,
seriousness, and types of later crime. Journal of Consulting and Clinical Psychology, 57,
710-718.
Stattin, H., & Magnusson, D. (1990). Pubertal maturation in female development. Hillsdale, NJ:
Erlbaum.
Stattin, H., & Magnusson, D. (1996). Antisocial behavior - a holistic perspective. Development
and Psychopathology, 8, 617-645.
Stattin, H., Magnusson, D., & Reichel, H. (1989). Criminal activity at different ages. A study
based on a Swedish longitudinal research population. British Journal of Criminology, 29,
368-385.
Steinberg, L., & Morris, A. S. (2000). Adolescent development. Annual Review of Psychology,
52, 83-110.
Stice, E., & Whitenton, K. (2002). Risk factors for body dissatisfaction in adolescent girls: A
longitudinal analysis. Developmental Psychology, 38, 669-678.
Stice, E., Presnell, K., & Bearman, S. K. (2001). Relation of early menarche to depression, eating
disorders, substance abuse, and comorbid psychopathology among adolescent girls.
Developmental Psychology, 37, 608-619.
76
36
Female
Pubertal Timing and Problem Behavior
Stone, C. P., & Barker, R. G. (1937). Aspects of personality and intelligence in postmenarcheal
and premenarcheal girls of the same chronological age. Journal of Comparative
Psychology, 23, 439-445.
Stone, C. P., & Barker, R. G. (1939). The attitudes and interests of premenarcheal and
postmenarcheal girls. Journal of Genetic Psychology, 54, 27-71.
Storvoll, E. E., Pape, H., & Rossow, I. (2008). Use of commercial and social sources of alcohol
by underage drinkers: The role of pubertal timing. Addictive Behaviors, 33, 161-166.
Susman, E. J., Dockray, S., Schiefelbein, V. L., Herwehe, S., Heaton, J. A., & Dorn, L. D.
(2007). Morningness/eveningness, morning-to-afternoon cortisol ratio, and antisocial
behavior problems during puberty. Developmental Psychology, 43, 811-822.
Susman, E. J., Dorn, L. D., & Schiefelbein, V. L. (2003). Puberty, sexuality, and health. In R. M.
Lerner & M. A. Easterbrooks (Eds.), Handbook of psychology: Developmental psychology,
vol. 6 (pp. 295-324). New York, NY: Wiley & Sons.
Susman, E.J. and A.D. Rogol (2004). Puberty and Psychological Development. In Lerner, R.M.
and L. Steinberg (Eds.), Handbook of Adolescent Psychology. New York: Wiley.
Tanner, J. M. (1989). Foetus into man: Physical growth from conception to maturity. Wale:
Castlemead Publications.
Tilton-Weaver, L. C., Vitunski, E. T., & Galambos, N. L. (2001). Five images of maturity in
adolescence: What does "grown up" mean? Journal of Adolescence, 24, 143-158.
Tschann, J. M., Adler, N. E., Irwin Jr., C. E., Millstein, S. G., Turner, R. A., & Kegeles, S. M.
(1994). Initiation of substance use in early adolescence: The roles of pubertal timing and
emotional distress. Health Psychology, 13, 326-333.
Turner, P. K., Runtz, M. G., & Galambos, N. L. (1999). Sexual abuse, pubertal timing, and
subjective age in adolescent girls: a research note. Journal of Reproductive and Infant
Psychology, 17, 111-118.
Udry, J. R. (1979). Age at menarche, at first intercourse, and at first pregnancy. Journal of
Biosocial Science, 11, 433-441.
Udry, J. (1988). Biological predispositions and social control in adolescent sexual behavior.
American Sociological Review, 53, 709-722.
Udry, J. R. (1990). Hormonal and social determinants of adolescent sexual initiation. In J.
Bancroft & J. M. Reinisch (Eds.), Adolescence and puberty. The Kinsey Institute series,
Vol. 3. (pp. 70-87). London, Oxford University Press.
Udry, J. R., & Billy, J. O. (1987). Initiation of coitus in early adolescence. American Sociological
Review, 52, 841-855.
Urberg, K. A., Degirmencioglu, S. M., & Pilgrim, C. (1997). Close friend and group influence on
adolescent cigarette smoking and alcohol use. Developmental Psychology, 33, 834-844.
Wadsworth, M. (1979). Roots of delinquency. Infancy, adolescence and crime. Oxford: Martin
Robertson.
Walker, E., Downety, G., & Bergman, A. (1989). The effects of parental psychopathology and
maltreatment on child behavior: a test of the diathesis- stress model. Child Development,
60, 15-25.
Westrin, P. A. (1967). Wit III Manual. Stockholm: Skandinaviska Testforlaget.
White, J. L., Moffitt, T. E., Caspi, A., Bartusch, D. J., Needles, D. J., & Stouthamer-Loeber, M.
(1994). Measuring impulsivity and examining its relationship to delinquency. Journal of
Abnormal Psychology, 103, 192-205.
37
Female Pubertal Timing and Problem Behavior
77
PAPER I
Wichstrom, L. (1999). The emergence of gender difference in depressed mood during
adolescence: The role of intensified gender socialization. Developmental Psychology, 35,
232-245.
Wichstrom, L. (2000). Psychological and behavioral factors unpredictive of disordered eating: A
prospective study of the general adolescent population in Norway. International Journal of
Eating Disorders, 28, 33-42.
Wichstrom, L. (2001). The impact of pubertal timing on adolescents’ alcohol use. Journal of
Research on Adolescence, 11, 131-150.
Wiesner, M., & Ittel, A. (2002). Relations of pubertal timing and depressive symptoms to
substance use in early adolescence. Journal of Early Adolescence, 22, 5-23.
Williams, J.M. & Currie, C.E. (2000). Self-esteem and physical development during early
adolescence: Pubertal timing and body image. Journal of Early Adolescence, 20, 129-149.
Wilson, D. M., Killen, J. D., Hayward, C., Robinson, T. N., Hammer, L. D., Kraemer, H. C.,
Varady, A., & Taylor, C. B. (1994). Timing and rate of sexual maturation and the onset of
cigarette and alcohol use among teenage girls. Archives of Pediatrics & Adolescent
Medicine, 148, 789-795.
Wolfson, S. (2000). Students’ estimates of the prevalence of drug use: evidence for a false
consensus effect. Psychology of Addictive Behaviors, 14, 295-298.
Zabin, L. (1984). The association between smoking and sexual behavior among teens in U.S.
contraceptive clinics. American Journal of Public Health, 74, 261-263.
Zehr, J. L., Culbert, K. M., Sisk, C. L., & Klump, K. L. (2007). An association of early puberty
with disordered eating and anxiety in a population of undergraduate women and men.
Hormones and Behavior, 52, 427-435.
Zelnik, M., Kantner, J. F., & Ford, K. (1981). Sex and pregnancy in adolescence. Beverly Hills:
Sage.
Zelnik, M., & Shaw, F. K. (1983). First intercourse among young Americans. Family Planning
Perspectives, 15, 64-70.
Zimmer-Gembeck, M.J. (2002). The development of romantic relationships and adaptations in
the system of peer relationships. Journal of Adolescent Health, 31, 216-225.
PAPER II
115
Klunserna Study
115
05-10-25, 10.22
1
PAPE R II
Adolescent and adult implications of girls’ pubertal timing: What
roles do perceived maturity and early sexuality play?
Therése Skoog
Center for Developmental Research
Department of Behavioral, Social, and Legal Sciences
Örebro University
Corresponding author:
Therése Skoog, Center for Developmental Research at BSR, Örebro University,
S-701 82 Örebro, Sweden
E-mail: [email protected]
Phone: +46 19 301091
Fax: +46 19 303484
1
2
”We talk a lot about how mature others at school are”
“Some at school are really childish and annoying. We say – they are so immature”
”When girls become mature, they can show who they really are”
Two 14 year-old Swedish girls when asked if maturity is important to them and their friends.
For adolescent girls, how mature they and their peers are seem to be of great concern. For
example, the girls who are quoted above talk in an interview about how mature they and their
friends are, and they repeatedly use the word “mature”. This is not very surprising,
considering that adolescence is a time of social comparison and maturation. Girls mature in
many ways. Biologically, they experience puberty, which results in reproductive capability
and an adult-like appearance. Psychologically, they start to think and feel differently about
themselves and the world around them. Socially, the characteristics of their social
relationships change. In adolescence, girls acquire new and different social roles with parents
and peers and many start having romantic and sexual relationships.
Girls’ developing bodies and minds
Adolescent girls mature notably in many different ways and their physical maturation is
doubtlessly the most obvious. Pubertal maturation in all domains is governed by hormones
(Traggiai & Stanhope, 2003). During this approximately four year long phase of anatomical
and physiological development, the sex organs mature and become functional. Some of the
changes are invisible to the eye, while others are apparent. A majority of the systems in the
body undergo changes during puberty (Tanner, 1978). Although girls experience the pubertal
changes in about the same sequence, they differ markedly with respect to how fast their
pubertal transitions are (Tanner, 1978).
The purpose of the development of the primary sexual characteristics – organs and
processes – is to make the individual capable of reproduction and to transform a sexually
immature girl into a sexually mature woman. Secondary sexual characteristics are not directly
associated with reproduction; instead they are responsible for the changes in appearance that
make girls look more similar to the adult female stereotype. In addition to the changes in
primary and secondary sexual characteristics, a rapid growth spurt takes place that gives
adolescents adult heights and weights (Traggiai & Stanhope, 2003). In sum, the aim of these
changes is to make a young girl’s body mature and attain reproductive capability, and to
signal this capacity to others.
Alongside the anatomical and physiological changes of puberty, girls experience
numerous psychological and social changes. Some of those changes are independent; others
are consequences, of puberty. A common belief among both scholars and lay people is that
going through puberty is challenging and troublesome. On the contrary, research in the latest
couple of decades has shown that puberty does not entail as much agony and trouble as was
previously thought. Most adolescents go through puberty without any particular problems.
Nevertheless, psychosocial adjustment difficulties increase drastically at the time around
puberty (e.g., Moffitt, 1993).
Pubertal status and pubertal timing
When social scientists examine the effects of puberty on girls’ behaviors, they either use girls’
pubertal status or girls’ pubertal timing as a basis for their analyses. Girls’ pubertal status
refers to where in the pubertal process girls are at a given point in time. One of the most
common ways to measure girls’ pubertal status is to use the so called Tanner stages (Tanner,
1978). Girls’ pubertal development is divided into five Tanner stages using girls’
development of secondary sexual characteristics for determining where in the pubertal
process girls are. The stages range from no signs of puberty, through menarche, to displaying
secondary sexual characteristics that resemble those of adult women (Tanner, 1978). Another
2
3
Girls’ puberty and psychosocial adjustment
How is puberty related to girls’ adjustment? A large body of research has shown that pubertal
timing matters for girls’ adjustment. According to the literature, girls who experience puberty
earlier than their same-age, same-sex peers tend have more social adjustment problems than
later-developing girls do (see Alsaker, 1995; Connolly, Paikoff, & Buchanan, 1996, for
reviews). The problems of early-developing girls range over many contexts. At home, they
have more disharmonious parent relations (Steinberg, 1987). During leisure time, early
developers are more delinquent, use more substances such as cigarettes, alcohol, and hashish,
and associate more often with delinquent and older peers (Deardroff, Gonzales, Christopher,
Roosa, & Millsap, 2005; Dubas, 2002; Flannery, Rowe, & Gulley, 1993; Graber, Lewinsohn,
Seeley, & Brooks-Gunn, 1997; Haynie, 2003; Tschann, Adler, Irwin, Millstein, Turner, &
Kegeles, 1994). Furthermore, they start engaging in sexual activities earlier than other girls
(Flannery et al., 1993). More problems at school have also been found to be related to
pubertal timing. Early-developing girls seem less motivated to achieve in school and they play
truant more often (Graber et al., 1997; Kaltiala-Heino, Marttunen, Rantanen, & Rimpela,
2003).
The main findings regarding girls’ pubertal timing and adolescent adjustment in
previous studies using the IDA sample (e.g., Stattin & Magnusson, 1990) are that earlydeveloping girls have more adjustment difficulties, are more sexually and romantically
advanced, and have a more unconventional peer network, which includes older and working
peers, compared to later-developing girls.
Taken together, there is much support in the literature that the adolescent transition of
early-developing girls is more troublesome than that of later-developing girls. The links
between pubertal timing and problem behaviour concern the group of early maturers as a
whole and they do not appear for every early maturing individual or even in every sample
(refs). This suggests that there must be important moderating factors. Furthermore, the links
are not of a great magnitude and most early maturing girls typically do not display severe
problem behavior. Because not all early maturing girls show adjustment difficulties, there
must be moderating factors that determine which early maturing girls will have problems and
which will not. Thus, it is only under certain circumstances that we should expect girls’
pubertal timing to be associated with adjustment difficulties.
Self-perceptions of maturity and heterosexual relationships in
adolescence
What other factors, then, might be important when explaining the role of pubertal timing in
female adjustment? It has been suggested that one reason why pubertal timing is more
important for girls’ adjustment than pubertal status is the importance of social comparisons in
adolescence (Alsaker, 1995). This suggests that girls’ cognitions and perceptions about
puberty may play important roles in the link between pubertal timing and adjustment. In fact,
3
PAPE R II
common method for determining girls’ pubertal status is to simply assess onset of menarche,
usually through self-reports.
The other basis for analyzing effects of individual differences in puberty is pubertal
timing. Although puberty is a universal transition and although girls generally experience the
pubertal events in about the same sequence, they differ markedly with respect to when they
experience puberty. Pubertal timing refers to when, in comparison with same-age, same-sex
peers or a standard norm, an individual experiences puberty. Today, some girls have their
menarche when they are 10 years old while others are 14 years old. The age range within
which puberty is considered normal is rather wide.
4
scholars have argued that it is when adolescents feel more mature that their pubertal timing
plays a meaningful role in their psychosocial adjustment (e.g., Alsaker, 1995). Consequently,
one potential factor that could differentiate meaningful subgroups of early developers, and
that could explain why only some early developers have more adjustment difficulties, is
whether girls feel that they are more mature than same-age peers or not.
Another potential factor that previously has been shown to play a role in the link
between pubertal timing and adjustment is early heterosexual relations. To understand how
the pubertal experience is related to girls’ adjustment, and to predict which girls are at risk for
developing problems during the adolescent period, researchers and scholars have forwarded
several hypotheses. Stattin and Magnusson (1990) proposed one such explanation of the link
between early pubertal timing and problem behavior. It is a social approach that builds on the
assumption that social comparisons are important and common in adolescence. Moreover, it
concerns girls’ perceptions about their maturity in relation to peers and early heterosexual
relations.
The hypothesis developed by Stattin and Magnusson (1990) predicts that individuals
who develop early and who engage with non-conventional peers, older and deviant peers, and
who have heterosexual relationships early, are at a higher risk of developing problem
behaviors. We build on this proposition in this chapter. We argue that due to their earlier
physical maturation, early-developing girls perceive themselves as psychologically and
socially more mature than their same-age peers do. Further, they are more interested in
”mature”, or adult-like, behaviors, in heterosexual contacts, and they identify with peers on
the ”same-maturity-level”. As a result, girls with early pubertal development will associate
with people who match their early developmental stance, through establishing heterosexual
relations, most often with older boyfriends, earlier than later-developing girls, and seeking out
and being sought out by peers who are chronologically older, but on the same maturity level.
Through relationships with older males and older peers, early-developing girls encounter
social environments with more advanced social behaviors, like drinking, drug use, and other
types of behaviors that are considered normbreaking for a 14 but not an 18-year-old, and are
introduced to leisure-time settings in which these types of behaviors are more typical. Earlydeveloping girls adjust to these social settings and peers that are more advanced, and start to
act and behave accordingly. Consequently, problem behavior among early-developing girls in
mid-adolescence may more be seen as engaging in socially more mature behavior for their
age, rather than as deviant behavior.
From this idea it follows that early developers who feel more mature than peers in
adolescence, and who have early romantic and sexual relations, should engage more in
socially advanced behaviors, including some normbreaking behaviors and drinking. In
contrast, there are few reasons to believe that early developers who do not feel mature and
who do not have romantic and sexual relations should have more adjustment difficulties than
others.
Maturity has many different meanings to adolescents, and physical maturity is just one
(Galambos, Barker, & Tilton-Weaver, 2003; Galambos, Kolaric, Sears, & Maggs, 1999;
Tilton-Weaver, Vitunski, & Galambos, 2001). Responsible behavior, power, and socially
advanced behaviors (drinking, having serious romantic partners, dressing sexy, etc.) are other
aspects of what adolescents see as signs of maturity (Tilton-Weaver et al., 2001). Hence, it is
possible that girls who perceive themselves as mature are more well-adjusted and adult-like in
positive respects, instead of being more normbreaking. This would be girls who study hard
and prepare for the future, and who take on responsibility for themselves and others, perhaps
to a greater extent than what would be expected for girls of their ages. Adolescents who feel
more mature than their age, and score high on assessments of psychosocial maturity and low
on normbreaking behavior have been found to be later developers. Conversely, girls who feel
4
5
Aim of this chapter
This study was designed to extend and refine previous work by Stattin and Magnusson (1990)
and others in two important respects. Given that early developers differ from others in terms
of adjustment, we need to examine the underlying developmental mechanisms. For this
purpose, we test the proposition that early pubertal timing have limited implications for
concurrent and future adjustment unless it is accompanied by greater perceived maturity and
early heterosexual relations. No known studies have examined implications of being an early
developer for adjustment in midlife. Therefore, we explore both young and middle adulthood
implications of pubertal timing. First, we examine whether girls with various constellations of
levels of pubertal timing, perceived maturity, and heterosexual relationships differ with regard
to psychosocial adjustment in mid-adolescence. Further, we investigate the implications of
these constellations in adolescence for women’s adjustment and lifestyle choices at age 26.
Finally, we explore the very long-term implications of these constellations in adolescence by
focusing on women’s adjustment at age 43. For all these analyses, the women in the IDA
sample with information on pubertal timing are included (n = 510; Magnusson, Dunér, &
Zetterblom, 1975).
5
PAPE R II
mature, but who score low on psychological maturity, and who have a lot of problem
behaviors are more likely to be early developers, and they are also more peer-oriented
(Galambos & Tilton-Weaver, 2000; Galambos et al., 2003). The latter group of girls has
therefore been referred to as “pseudomature” (Galambos & Tilton-Weaver, 2000; Galambos
et al., 2003). In sum, the reasons for feeling mature could be different for different
individuals, and feeling mature could have different implications depending on the girls’
reasons for doing so. According to these previous findings, (Galambos & Tilton-Weaver,
2000; Galambos et al., 2003), early developers who feel mature should be at greater risk for
adjustment problems than late developers who feel mature.
In addition to perceived maturity, Stattin and Magnusson’s (1990) explanation of earlydeveloping girls’ tendency for problem behaviors include heterosexual relationships in midadolescence. Romantic relationships and sexuality are major issues for most adolescents. As
girls go through adolescence, they become more romantically involved and sexually
experienced. By late adolescence, the majority of girls have experienced sexual intercourse
(Edgardh, 2000). However, there are considerable differences in the timing of sexual onset.
A body of literature has shown that early age at first intercourse is linked with many
concurrent and future adjustment problems, such as early pregnancy (Fergusson, Horwood, &
Lynskey, 1994; O’Donnell, O’Donnell, & Stueve, 2001; Orr, Beiter, & Ingersoll, 1991).
Deardorff and colleagues (2005) found that early sexual intercourse mediates the link between
early pubertal timing and having children at an early age. This implies that the reason why
early developers become pregnant earlier than other girls do is because they have early sexual
intercourse. Girls who are engaged in romantic relationships tend to have boyfriends that are
somewhat older then the girls are themselves (Marin, Coyle, Gomez, Carvajal, & Kirby,
2000). Because the boyfriends are chronologically older, they will have more socially
advanced behaviors, such as alcohol drinking. Hence, there the causal chain might be that
early-developing girls feel more mature and therefore associate with older peers and
boyfriends, who have a more socially advanced behavior. Through those older boyfriends
early developers are canalized into more advanced behaviors, which are considered
normbreaking for girls their age.
6
Method
Participants
The sample came from the longitudinal research program Individual Development and
Adaptation (IDA). The participants come from a medium-sized Swedish town - Örebro. At
the outset of the investigation in 1965, the city of Örebro included about 80,000 inhabitants
and has expanded over time. In 1985 there were approximately 120,000 persons living in
Örebro and in year 2000 130,000 persons. Given the population of Sweden (currently 9
million), Örebro is a relatively large city by national standards. The economy of Örebro was
at the initiation of the study dominated by a prominent footwear industry, but has diversified
over time into engineering plants, printing shops, and a food and paper industry. Additionally,
Örebro is home to a large hospital and university. A city-based comparison (Stattin,
Magnusson, & Reichel, 1986) showed that Örebro had comparable rates of crime to similarsized cities in Sweden (50,000-100,000 inhabitants), and moderately lower rates of crime
compared to Sweden’s largest cities (i.e., Stockholm, Göteborg, Malmö).
The first data was collected when the participants were 10 years old and data gathering
is still in progress. The target sample comprised all children who in 1965 attended normal
grade 3 schooling in public compulsory school. There were no private schools in the
community at this time. The total number of children was 1,027. Of those, 510 were girls and
517 boys. Data from those 510 girls were used in Papers II and III of this dissertation.
Thus far, data has been collected at ages 10, 13, 15, 16, 17, 18, 27, and 43 for the female
participants. The sample size has grown bigger over time. This is because girls who moved
into the community entered the study while girls who left the community did not leave the
study. At the last data collection, in 1998, 682 women participated. Information has been
obtained through official registers, self-reports, peer and teacher ratings, interviews, postal
questionnaires, and medical information. Parent reports have also been included. At age 27,
participants filled out postal questionnaires containing questions on childbirth, education,
work, and drinking habits. Of those adolescents who participated in adolescence, 90% took
part in the follow-up at age 27. When the women were 43 years of age, in 1998, they were
interviewed about various aspects of their adjustment situations. The participation rate was
89%.
Measures
We use age at menarche as our measure of girls’ pubertal timing. Age at menarche has been
considered the most salient indicator of pubertal development for girls (Graber, Petersen, &
Brooks-Gunn, 1996). It is also the most widely used measure of pubertal development in girls
(Brooks-Gunn & Graber, 1994). Using the IDA sample, Stattin and Magnusson (1990)
compared this self-report measure with bone ossification and weight and height measures at
the age of 13 years. They found strong support of validity for the self-reported measure of age
at menarche as a measure of girls’ pubertal timing.
The girls were asked about their ages at menarche in grade 8 (the age of 15). Five girls
(1%) had their first menstruation before the age of 10, 46 (9%) between 10 and 11 years, 117
(22.9%) between 11 and 12 years, 207 (40.6%) between 12 and 13, 106 (20.8%) at the age of
13 or later, and 29 girls (5.7%) had not had their first menstruation at the time of the data
collection. Thus, the typical age of menarche among the girls in the sample was between 12
and 13 years. In addition to this continuous measure, we created a categorical measure. Girls,
who had their menarche before age 12, were assigned to an early developers group. Girls,
who had their menarche after age 12, were assigned to a later developers group.
Second, we were interested in girls’ perceptions of their maturity or how mature they feel.
In grade 8, in a questionnaire that measured different aspects of peer and parent relations, the
girls were asked if they felt more or less mature than their classmates, on a 5-point scale with
6
7
Subgroups and statistical analyses
We divided the girls into eight subgroups based on the dichotomous measures of pubertal
timing, perceived maturity, and heterosexual relations. The first group consisted of the laterdeveloping girls who did not feel more mature than peers, and who had little of heterosexual
relationships (the Reference Group; n = 131). The second group contained the laterdeveloping girls who did not feel more mature than peers, and who had much of heterosexual
relationships (the Late Relationship-Only Group; n = 87). The third group consisted of the
later-developing girls who felt more mature than peers, and who had little of heterosexual
relationships (the Late Mature-only Group; n = 22). The fourth group contained the laterdeveloping girls who felt more mature than peers, and who had much of heterosexual
relationships (the Late Precocious Group; n = 30). The fifth group consisted of the earlydeveloping girls who did not feel more mature than peers, and who had little of heterosexual
relationships (the Premature Group; n = 32). The sixth group consisted of the earlydeveloping girls who did not feel more mature than peers, and who had much of heterosexual
relationships (the Early Relationship-Only Group; n = 50). The seventh group consisted of the
early-developing girls who felt more mature than peers, and who had who had little of
heterosexual relationships (the Early Mature-Only Group; n = 13). The eighth group consisted
of the early-developing girls who felt more mature than peers, and who had much of
heterosexual relationships (the Early Precocious Group; n = 41). Partial drop-out occurred,
making the groups somewhat smaller and different between tables.
We employed one-way analyses of variance to compare the subgroups on the
adjustment measures. Post hoc analyses were conducted using the LSD procedure. We also
performed a configural frequency analysis (CFA) using the statistical package SLEIPNER to
test developmental trajectories (Bergman & El-Khouri, 2002; Lienert & zur Oeveste, 1985).
Expected and observed frequencies for each configuration were compared. An outcome was
labeled “typical” when the observed frequency in a cell exceeded the expected frequency. In
reverse, when the expected frequency in a cell exceeded the observed frequency, the outcome
was labeled “atypical”.
7
PAPE R II
the scale values (1) not at all as mature as most others; (2) not really as mature as most
others; (3) about as mature as most others; (4) somewhat more mature; and (5) much more
mature than most others. In addition to this continuous measure, we created a dichotomized
measure dividing the girls into groups of girls who felt more mature than their peers (scores 4
and 5) and girls who did not feel more mature than their peers (scores 1, 2, and 3).
Third, we were interested in girls’ heterosexual relations. Four items measured how
advanced the girls’ heterosexual relationships to boys were in grade 8. Girls were asked about
their perceived sexual maturation, “Do you feel sexually more experienced than your agemates? (1=never, 2=seldom, 3=sometimes, 4=rather often, and 5=very often), intercourse,
“Have you had sexual intercourse?” (1=no, 2=yes, once, and 3=yes, several times), attitude
towards intercourse (1=very stupid to 7=totally OK), and steady boyfriend relations “Do you
have or have you had a steady relation to a boy?” (1=have never had, don’t want to have,
2=have never had, but I want now, 3=have had earlier, but not now, 4=have now, but not
earlier, and 5=have now, and have had earlier). Since the items had different numbers of
response alternatives, we standardized them before computing the composite scale. The alpha
reliability for this scale was .73. The girls were then divided into a group with little of
heterosexual relationships and a group with much heterosexual relationships using a median
split.
Outcome measures are described in connection with the results.
8
Results
Adjustment in adolescence
Concerning adolescence, we examined girls’ adjustment at home, in school, and during
leisure time. We have few reasons to believe that the later-developing girls should have high
levels of adjustment difficulties. In contrast, those girls should be among the well-adjusted,
particularly if they belong to the Late Mature-only group. Given that early-developing girls
have to feel more mature for the link between pubertal timing and psychosocial adjustment
difficulties to exist as proposed above, it follows that the groups with early-developing girls
who do not feel mature should have less adjustment problems than the groups with girls who
are early-developing and who feel mature. The Early Precocious group should have the most
adjustment difficulties. Table 1 reports correlations between pubertal timing, perceived
maturity, and heterosexual relationships on the one hand and all adjustment variables in
adolescence on the other (using the continuous measures). Table 2 presents results of one-way
analyses of variance with the eight groups as the predictor variable and the social adjustment
variables as outcome variables.
Adjustment at home – Relationships with parents. During puberty, girls
become increasingly interested in making their own decisions and act on their own,
independent of parents. It is likely that this could create conflicts and turmoil in the family.
We were interested in whether girls in the different groups differed in terms of relationships
with parents. We employed a scale comprising 7 items to measure problematic parent
relations. Examples of items were: “Do your parents listen to you?” “Are your parents
disappointed with you?” The scale was internally consistent (a = .87).
Pubertal timing and heterosexual relations, but not perceived maturity, was related to
having problematic parent relations. Moreover, the groups differed in terms of having
problematic parent relations. A post hoc analysis showed that the early precocious group
differed significantly from all other groups in terms of having more problematic parent
relations. In stark contrast, the Early Mature-Only group had the least problematic parent
relationships of all groups. This indicates that early pubertal timing mean very different things
for girls’ relations to parents, and depends on whether girls feel mature and have early
heterosexual relations.
School adjustment. Adolescents spend a lot of their time in school. As such, how
they feel about school and their teachers is of great interest. Moreover, it is likely that how
girls feel at school when they are young influences whether they stay in school in the long
run. We examined girls’ adjustment to school in grade 8. This was measured with a scale
comprising nine items. Some of the items were: “Are you satisfied with your school work?”
“Is it important to perform well in school?” “Are your teachers fair to you?” The alpha
reliability for this scale was .83.
There was a significant correlation between pubertal timing, and heterosexual
relationships and school adjustment, but not between perceived maturity and school
adjustment. The more of early developers the girls were, and the more heterosexual
relationships they had, the worse they adjusted to school.
Girls in the eight groups differed significantly in terms of their school adjustment. The
Early Precocious and the Early Relationship-Only groups had the worst adjustment to school
among the groups of girls. The late Mature-Only group had the least adjustment difficulties
in school. According to the post hoc analysis, this group differed significantly from the Late
Relationship-Only and the Late Precocious groups.
Leisure time adjustment. Peer relationships. As girls proceed through adolescence,
peer relationships become increasingly important. The size of peer networks increases and
8
9
9
PAPE R II
girls spend more time with peers. Are girls’ pubertal timing, perception of maturity, and
heterosexual relationships associated with peer relationships in adolescence?
We explored how many peers girls had of different sorts and how often girls associated
with peers in grade 8. Girls reported the number of peers they had and how many days per
week they met with their peers (peer regularity). Girls’ pubertal timing, perceived maturity,
and heterosexual relationships all correlated with total number of peers, number of older
peers, and number of working peers, such that earlier developing girls, girls who felt more
mature, and girls who had more heterosexual relationships had more of those types of peers.
Only heterosexual relationships correlated with how often girls met with their peers, where
the more of heterosexual relationships girls had, the more often they met with their peers.
When we examined the constellations of pubertal timing, perceived maturity, and
heterosexual relations, we found that the Early Precocious group had more peers, more older
and working peers, than the other groups. A post hoc analysis revealed that they deviated
from the Premature group in terms of having older and working peers. In terms of the total
number of peers, the Early Precocious group differed from the premature group significantly.
In addition, the groups differed in terms of how often they associated with their peers.
Again, the Early Precocious group associated the most with their peers. A post hoc analysis
indicated that they only differed from the Late Mature-Only group though.
Normbreaking. Adolescence is a period of increased problem behavior, including
substance use and delinquency (e.g., Moffitt, 1993). Did the eight groups differ in terms of
norm breaking behavior?
Girls’ normbreaking was a scale aggregated from eleven questions about shop lifting,
signature forgery, pilfering, using hashish, vandalism, evasion of payment, physical
aggression, harassment, and loitering that were asked in grade 8 (Magnusson et al., 1975).
The 5-point response scale had the following options: 1=no, 2=once, 3=2-3 times, 4=4-10
times, and 5=more than 10 times. The alpha reliability for the scale was .73.
Pubertal timing, perceived maturity, and heterosexual relationships were all correlated
with normbreaking behavior; the more of early developers the girls were, the more mature
they felt, and the more heterosexual relationships they had, the more normbreaking behavior
they engaged in. The eight groups also differed from each other in terms of normbreaking
behavior. A post hoc analysis revealed that the Early Precocious group differed significantly
from all other groups. They had higher levels of normbreaking than any other group of girls.
In contrast, the Premature and the Early Mature-Only group had among the lowest levels of
normbreaking behavior compared to the other groups.
Alcohol drinking. We wanted to explore girls’ drinking separately from other types of
normbreaking behaviors, because of the heightened awareness of adolescent girls’ drinking
behavior. Girls were asked whether they had ever been drunk. The response scale was the
same as for the items tapping normbreaking behavior. All three predictor variables were
linked with intoxication. Moreover, the groups differed significantly in terms of intoxication.
Again, a post hoc analysis revealed that the Early Precocious group differed significantly
from all other groups of girls in terms of having been drunk more often. The Late MatureOnly group reported being drunk most seldom.
10
Table 1. Correlations between pubertal timing, perceived maturity and heterosexual
relationships, and adjustment measures in adolescence
Pubertal timing
.30***
.29***
.18***
Perceived maturity
Heterosexual
relationships
.25***
.05
.29***
.16***
.11*
.24***
.22***
.09
.23***
.25***
-.01
.15**
.26***
.14**
.02
.14***
.15***
.30***
.20***
.31***
.30***
.16**
.44***
.61***
Perceived maturity
Heterosexual relationships
Problematic parent
relationships
School maladjustment
Number of peers
Older peers
Working peers
Peer regularity
Normbreaking
Frequency of drunkenness
* p<.05 ** p<.01 *** p<.001
10
.22
-.11
4.16
1.10
.13
3.60
-.18
1.54
Number of peers
Older peers
Working peers
Peer regularity
Normbreaking
Frequency of
drunkenness
3.22
3.94
.20
.85
4.90
2.08
.19
The Premature group
.01
The Early relationship-Only
group
Problematic parent
relationships
School adjustment
The Early Mature-Only group
1.50
3.23
-.29
.62
3.85
1.38
-.21
-.45
The Early Precocious group
3.55
4.03
.51
1.10
5.79
3.31
.22
.43
The Reference group
1.60
3.57
-.18
.15
4.13
.96
-.11
-.16
The Late Relationship-Only
group
2.71
3.69
.05
.45
4.75
1.30
.05
.08
The Late Mature-Only group
1.37
3.22
-.28
.50
4.79
1.82
-.39
-.28
3.28
3.59
.15
.73
5.27
2.48
.18
.14
The Late Precocious group
7,391
7,406
7,467
7,420
7,420
7,420
7,467
7,467
28.49***
2.65**
13.06***
6.27***
2.69**
8.05***
5.09***
6.06***
1>3; 2,6,8 > 3,5,7;
4>1,3,5,6,7
2,4>1,3,5,7; 6>5,7;
8>3,5,7; 5>7
4>1,3,5,6; 8>5
2,8>1,5,6;
4>1,2,3,5,6,7; 7>5
2>1,5,6; 4>1,5,6,7;
6>5; 8>1,5
2,4>3,5,7; 6>7
2,6,8>1,3,5,7;
4>1,2,3,5,6,7,8
2,8>1,3,5,6,7;
4>1,2,3,5,6,7;
6>1,3,5,7;
Table 2. Eight groups of girls (puberty, perceived maturation, heterosexual relationships) compared on adjustment in adolescence (grade 8).
Groups of girls
df
F
Post Hoc
PAPE R II
11
Conclusions about the link between pubertal maturation and
adolescent adjustment
It appears as if pubertal timing plays a significant role in girls’ adjustment in
adolescence. This appears also to be the case with perceived maturity and heterosexual
relations. However, the findings clearly indicate that we get a fuller understanding of girls’
adolescent adjustment when combining these two aspects of maturity and heterosexual
relationships, and examine implications of them simultaneously. Overall, girls who were later
developers did not have problematic adjustment in adolescence. Interestingly, the Early
Mature-Only group did not have any particular adjustment difficulties either. In fact, this
group had the lowest levels of several of the problem indicators in adolescence, including
normbreaking behavior. By contrast, the Early Precocious group had more problems with
parents, had more peer interactions, including older and working peers, had more schoolrelated problems, were more heterosexually advanced, had more often been drunk, and were
more normbreaking compared to the other girls. Taken together, the findings concerning the
adolescent period suggest that it is the constellation of pubertal timing, perceived maturity,
and heterosexual relations, not pubertal timing, perceived maturity, and heterosexual
relationships seen separately, that are important for girls’ adjustment in adolescence.
Since girls’ patterns of maturity and opposite-sex relationships seem to play significant
roles in their adolescent adjustment, it is plausible that this continues to be important for the
girls’ adjustment later in life, as adult women.
Adjustment in young adulthood
Two main domains in young women’s lives are family and work. Although most
women plan to have children sometime in the future, young women’s main commitment often
seems to be either their families or their educations and professional careers (O’Laughlin &
Anderson, 2001). As a consequence, the timing of when women choose to marry and have
their children varies. Whether women choose to aspire for high educational attainments and
professional careers also varies markedly. Career-oriented women tend to delay childbearing
by several years (Wu & MacNeill, 2002). This indicates that women seem to be either family
or career oriented in young adulthood.
Apart from that early developers are more likely than later developers to become
pregnant and have children early (Deardorff et al., 2005, Hockaday, Crase, Shelley, &
Stockdale, 2000; Manlove, 1997; Stattin & Magnusson, 1990; Woodward, Fergusson, &
Horwood, 2001), the information about adult implications of girls’ pubertal timing is
somewhat inconclusive. Dubas (2002) found that early developers were less likely to suffer
from externalizing problems, including delinquency and soft drug use, than later-developing
women were in their early twenties. In contrast, Graber and colleagues (2004) found an
association between adult psychopathology and female pubertal timing. At 24 years, early
developers had had more disruptive behaviors, antisocial personality disorders, and
internalizing problems than late-developing females over their life courses. However, the
problems of the early-developing women largely arose during the adolescent years, rather
than in early adulthood. Moreover, they found no differences between early- and latedeveloping women in terms of level of education. Furthermore, in one of the most thorough
studies of long-term consequences of pubertal timing, Stattin and Magnusson (1990) found
that early-developing women in the IDA sample were more likely to have children, and a
larger number of children, than later-developing women did at age 26. In addition, early
developers had lower educational attainments and work positions than did the other women.
Thus, in spite of the lack of conclusive findings about adult implications of girls’
pubertal timing, studies indicate that early-developing girls as a group are at risk. However, as
in adolescence there might be subgroups of women within the group of early developers for
12
Family formation
What should we expect to find in terms of family formation of the eight groups of
women in young adulthood? Perhaps we should expect early developers to become mothers
earlier than the late developers. The reasoning behind this prediction is simply that early
developers have had the chance to get pregnant for a longer period of time than other women
have in young adulthood. Early sexual onset is predictive of early motherhood in young
adulthood (O’Donnell et al., 2001). Moreover, Deardorff and colleagues (2005) found that
early sexual relations mediated the link between early pubertal timing and early motherhood.
Therefore, we could expect more of the early developers who had early heterosexual
relationships to have children early, compared to other young women.
There is some literature on the meaning of perceived maturity in adolescence (Galambos
& Tilton-Weaver, 2000; Galambos et al., 2003). But the long-term consequences of selfperceived maturity in the teenage years are not yet clear. Therefore, we can make no clear-cut
predictions about what implications perceived maturity in adolescence will have for women’s
future adjustment. However, the Early Precocious group had socially advanced behaviors in
adolescence (such as associating with older peers and drinking) so it is possible that they
reach adult status in some aspects before others of the same age do. Hence, we should expect
those girls to have a faster transition into adulthood, which should be characterized by early
motherhood.
Marital status. First, we investigated whether the eight groups of women differed in
terms of their marital status. They were asked if they were married/cohabitants or not. A Chi2
test showed no differences between the groups in terms of whether women were
married/cohabitants or not.
Motherhood at age 26. Participants were asked how many children they had. Second,
participants reported on the age of their oldest child. Pubertal timing and early heterosexual
relationships were correlated with number of children. The eight groups of women differed
significantly with respect to how many children they had. The Early Precocious group had the
highest number of children. On the other hand, the Early Mature-Only group did not differ
from other women in terms of how many children they had. The Late Mature-Only group had
the lowest number of children.
Pubertal timing and heterosexual relationships were correlated with the ages of the
women’s oldest child. Of those who had children, the Early Precocious group had the oldest
first-borns. However, because the groups were so small, the difference was only approaching
significance (p = .08).
13
PAPE R II
whom these findings do not apply. To achieve a more comprehensive picture of adult
implications of girls’ pubertal timing, we will investigate various constellations of pubertal
timing, perceived maturity in mid-adolescence, and early heterosexual relationships and their
associations with family formation, career development, and criminality at age 26. We use the
same eight groups as we did when investigating adolescent adjustment.
Table 3 reports correlations between pubertal timing, perceived maturity, and heterosexual
relationships on the one hand, and adjustment variables in young adulthood on the other.
Table 4 presents results of one-way analyses of variance with the eight groups as the predictor
variable and the social adjustment variables in young adulthood as outcome variables.
Educational and career achievements
Because the Early Precocious group had the worst school adjustment in adolescence, we
could expect them to shy away from school. Further, if there is a conflict between committing
either to a family or to education and a professional career, early-developing girls should be
expected to have lower aspirations for education and career. In line with this, early
heterosexual debut has been found to be predictive of low educational attainments in young
adulthood (O’Donnell et al., 2001). Consequently, early-developing girls might have lower
educational attainments as young adults. The Late Mature-Only group, on the other hand,
should be prone to stay in academia for a longer time and hence have higher educational
achievements in young adulthood.
Females reported on their level of education on an 8-point scale, ranging from no
education past compulsory school (ending at age 16) to at least four years of university
studies. Pubertal timing and heterosexual relationships correlated with women’s educational
levels. There were significant differences between the eight groups of women with respect to
their educational achievements. The Late Mature-Only group had the highest educational
attainments. The Early Precocious group had the lowest. The Premature group did not differ
much from other women in terms of their educational levels.
In addition, we asked whether participants had employments and they answered an open
question about their types of employment at 26 years. From this information, job descriptions
were coded on an 8-point work position scale according to the minimum education required
for each kind of employment. Pubertal timing and heterosexual relationships were both
associated with women’s work positions, where early pubertal timing and much of
heterosexual relationships were associated with having lower work positions. Women in the
eight groups differed in terms of their work positions in a similar way as for their educational
attainments. The Late Mature-Only group had work positions that required the highest
educational levels. The Early Precocious group had the lowest work positions. The Premature
group again did not differ from other women in terms of their work positions.
Alcohol consumption
The Early Precocious group had developed socially advanced behaviors in adolescence
to a greater extent than other girls. One of these behaviors were alcohol drinking. Because of
our findings on the girls’ adjustment in adolescence, it is possible that the early precocious
group drink more later on in life too. We base this assumption on the established association
between early alcohol debut and later heavy drinking in prior studies (e.g., Pitkanen, Lyyra, &
Pulkkinen, 2005).
The women were asked about their alcohol drinking habits. They reported on how much
alcohol they drank at the most during one evening.
Early heterosexual relationships were linked with women’s alcohol consumption in
young adulthood, such that the more heterosexual relationships girls had, the more they drank
in young adulthood. There was a small difference between the women in the eight groups in
terms of their alcohol consumption. The Late Relationship-Only group had the highest
consumption levels. Hence, alcohol consumption seemed to be independent of pubertal
maturation, both actual and perceived.
Criminality
In adolescence, the Early Precocious group had higher levels of normbreaking than all
other groups. One could therefore expect this group to be overrepresented among those who
are high on criminality in young adulthood too. On the other hand, it is possible that their
normbreaking behavior is in effect adolescent limited (see Moffitt, 1993). Information about
criminal offences was obtained from official registers. For underage offending (below the age
14
Table 3. Correlations between pubertal timing and perceived maturity and adjustment
measures in young adulthood
Pubertal timing Perceived maturity
Number of children
.18***
.01
Age of oldest child
.20†
.13
Level of education
-.13**
.01
Work position
-.10*
-.01
Alcohol consumption
.02
.01
Criminality
.05
.03
†
p<.10 * p<.05 ** p<.01 *** p<.001
Heterosexual relationships
.18***
.37***
-.20***
-.11*
.19***
.11*
15
PAPE R II
of 15), information about juvenile offending was obtained from local welfare authorities.
Information about registered offences between the age of 15 and the age of 30 were obtained
from police registers. Only 3 of the girls had underage offending. At the age of 30, 7.5% of
the women with data on pubertal timing and perceived maturity were registered by the police
for one or several criminal offences.
Early heterosexual relationships were correlated with criminality, such that the more
heterosexual relationships girls had in mid-adolescence, the more criminal offences they had
committed in young adulthood. Women in the eight groups did not differ in terms of their
levels of lawbreaking though. Hence, criminality in young adulthood seemed to be
independent of actual and perceived pubertal maturation.
56.03
.17
2.74
47.88
.08
Work position
Alcohol
consumption
Criminality
2.43
2.71
3.27
1.22
3.88
.86
Age of oldest child
Level of education
The Premature group
.71
The Early relationship-Only
group
Number of children
The Early Mature-Only group
.06
46.56
2.13
1.00
3.69
.38
The Early Precocious group
.07
58.18
1.88
3.90
3.12
.93
The Reference group
.04
46.06
2.41
.91
3.75
.47
The Late Relationship-Only
group
.07
66.41
2.53
1.78
3.34
.58
.07
43.26
3.81
1.00
4.85
.19
The Late Mature-Only group
.06
57.96
2.39
2.10
4.23
.32
7,467
7,329
7,419
7,75
7,418
7,417
1.51 n.s.
4.48*** 1>7,8; 2>3,5,6,7,8;
4>3,5,6,7,8; 6<7
1.94†
2>5; 4>1,5,7
4.55*** 4<1,5;
7>1,2,3,4,5,6;
8>2,4,6
2.64** 7>1,2,3,4,5,6,8;
1>4
2.57** 6>1,5,7
Table 4. Eight groups of women (puberty, perceived maturation, heterosexual relationships) compared on adjustment in young adulthood (age
26).
Groups of girls
df
F
Post Hoc
The Late Precocious group
16
Developmental paths
17
PAPE R II
Because the early precocious group had a faster transition to motherhood in young
adulthood, we wanted to test the idea that some early-developing girls were set on a
developmental trajectory from feeling mature and early heterosexual contacts in adolescence
to early family formation in young adulthood.
For this purpose, we performed a configural frequency analysis (Lienert & zur Oeveste,
1985). If early developers are on a trajectory involving feeling mature, having much of early
heterosexual relationships in adolescence and early family formation in adulthood, then a
pattern of developing early, feeling mature, having much of early heterosexual relations, and
having children in young adulthood should constitute a “type” (i.e., should occur more
frequently than expected by chance).
Table 5 reports the findings of the trajectory analysis. It indicates that it was typical of
the Early Precocious and the Early Relationship-Only group to be mothers in young
adulthood. The trajectory involving the Premature group who did not have children early was
atypical (i.e., occurred less frequently than expected by chance). It was not typical for the
Premature and the Early Mature-Only group to be mothers in young adulthood. These
findings support the claim that girls’ pubertal timing has to be viewed in relation to other
circumstances in order to make predictions of which girls will be mothers early in life.
18
Later Precocious
Later Mature-Only
Late Relationship-Only
Reference group
Early Precocious
Early Mature-Only
Early Relationship-Only
Premature
Group
Actual
Pubertal
Maturation
Early Late
Perceived
Maturation
Low High
Adolescence
Level of
Heterosexual
Relationships
Low High
Children
No Yes
Young
adulthood
Obs. f
20
15
16
35
11
5
18
24
90
49
49
36
22
4
23
8
Exp. f
31.28
22.11
30.26
21.39
11.60
8.20
11.23
7.94
61.03
43.14
59.05
41.74
22.64
16
21.91
15.48
Chi2
4.07
2.28
6.72
8.66
.03
1.25
4.09
32.52
13.75
.80
1.71
.79
.02
9.00
.05
3.62
p
.02
.07
.00
.00
.50
.17
.04
.00
.00
.19
.09
.20
.50
.00
.44
.03
Antitype
Antitype
Type
Type
Antitype
Type
Antitype
Table 5. Results of a configural frequency analysis of pubertal timing, perceived maturity, and heterosexual relationships in mid-adolescence, and
motherhood at age 26.
Conclusions about young adulthood adjustment
Adjustment in midlife
We now turn to women’s adjustment at age 43. Does women’s pubertal timing still have
implications for their adjustment situation here in mid-life, thirty years after they had their
first menstruation? At this age, the IDA women were asked how old they were when they
experienced menarche. This measure correlated highly with the same measure collected when
girls were in grade 8, nearly 30 years earlier (r = .69, p < .001). This suggests that even in
midlife, women accurately remember when they experienced puberty.
Almost no previous published work has examined the very long-term implications of
girls’ pubertal timing on their adjustment. In one study, Johansson and Ritzén (2005) used the
IDA sample to investigate effects of girls’ pubertal timing on some aspects of women’s
physical status at age 43. The findings indicated that early developers were of shorter and
heavier stature; they had a higher body mass index and they were less physically fit. In
addition, early developers reported more frequent dieting than the other women did. Also the
women’s life satisfaction and partner relationship quality at age 43 were examined. No effects
of women’s pubertal timing on either their life satisfaction or their partner relationship quality
at age 43 were found. In the present chapter, we investigate women’s adjustment with respect
to family life and education. We use the same eight groups as we did for investigating
adjustment in adolescent and young adulthood. Table 6 presents results of one-way analyses
of variance with the eight groups as the predictor variable and romantic partner, motherhood,
and educational level as outcome variables.
Family life
Motherhood at age 43. The women reported on number of children they had given
birth to by age 43. This measure includes children living with the participant and children
who had already left their parental homes. No differences were found between the groups on
this measure.
Romantic partner. The women reported on whether they lived with a husband or a
cohabitant. There were no differences between the eight groups in terms of living with a
romantic partner.
19
PAPE R II
As for the adolescent period, the combination of pubertal timing, perceived maturity,
and heterosexual relationships seems to be a fruitful predictor of some aspects of women’s
adjustment in young adulthood. A general pattern that emerged was that the Early Precocious
group was oriented toward family formation rather than professional careers. Interestingly, for
the Premature group, this was not equally true. Women who felt mature in mid-adolescence,
but who experienced puberty relatively late, were more oriented toward achieving high
educational levels and less oriented towards family formation. This was particularly the case
when these later developers had little of heterosexual relationships in mid-adolescence.
Women’s criminality and alcohol consumption, on the other hand, seemed to have little to do
with their pubertal timing and perceived maturity in adolescence. Rather, early heterosexual
relationships were linked with higher alcohol consumption and criminality.
Trajectory analyses showed that girls who experienced puberty early and who felt more
mature than peers appeared to be set on a developmental pathway of having heterosexual
relationships early and an early family formation in young adulthood. Perhaps most
interesting was that the pattern of the Early Premature group who had children in young
adulthood was not typical, while the pattern of the Early Precocious group who had children
in young adulthood was highly typical. These findings demonstrate the importance of taking
adolescent girls’ perceptions of how mature they are and their heterosexual relationships into
consideration when examining long-term implications of pubertal timing.
Education
Own educational level. Previous research has shown that there are no differences
between early and later developers in the IDA sample in terms of their educational
achievements in midlife (Johansson & Ritzén, 2005). The question is whether taking
perceived maturity and early heterosexual relationships into account changes the picture. At
age 43, females responded to a question about highest educational level. Answers were coded
on a 7-point scale ranging from no education past compulsory school, to at least a university
degree. There was also an “other” answer option. This option was omitted from the analyses.
Women’s educational levels were stable between young adulthood and midlife (r = .69,
p < 001). Results indicated that there were differences in educational levels between the
groups. The Late Mature-Only group had the highest educational achievements in mid-life.
The other groups were similar in terms of educational levels.
Partner’s educational level. The women reported on their partners’ highest educational
levels on 10-point scales, were 10 indicated the highest education level (a university degree).
Results showed that the eight groups did differ and a post hoc revealed that the Late MatureOnly group had partners with the highest educational levels. Early developers who had much
of early heterosexual relationships had partners with the lowest educational achievements.
20
Number of children
Living with a romantic
partner
Own level of education
Partner’s level of
education
4.02
5.33
4.22
6.10
The Premature group
1.98
.78
The Early relationship-Only
group
1.81
.78
The Early Mature-Only group
4.77
6.11
1.92
.77
The Early Precocious group
4.78
5.81
2.15
.88
The Reference group
4.55
6.69
1.94
.81
The Late Relationship-Only
group
3.95
5.62
1.92
.75
The Late Mature-Only group
5.68
8.63
2.05
.91
5.74
7.81
2.00
.93
The Late Precocious group
7,400
7,290
7,398
7,398
3.14** 7,8>1,2,5,6
4.05*** 5>2,6; 7>1,2,3,4,5,6;
8>2,4,6
0.26 n.s.
1.21 n.s.
Table 6. Eight groups of women (puberty, perceived maturation, heterosexual relationships) compared on adjustment in midlife (age 43).
Groups of girls
df
F
Post Hoc
PAPE R II
21
Conclusions about midlife findings
Besides the findings that previously have been reported regarding effects on physical
status (Johansson & Ritzén, 2005), there seems to be few differences between the sub-groups
concerning adjustment in midlife, at least in terms of the aspects of adjustment that we
examine here. In terms of childbearing, for instance, it appears as if the other groups catch up
on the group of early developers who felt mature in adolescence and who had their first child
earlier in life. When using pubertal timing as the sole predictor, there is no difference between
women in terms of their educational levels at age 43 (Johansson & Ritzén, 2005). However,
when we examined pubertal timing and perceived maturity simultaneously in this study, later
developers who felt mature with or without early heterosexual relationships had significantly
higher educational attainments than the other women. In addition, those women had partners
with the highest educational levels. Thus, there seemed to be some, although few,
implications of girls’ belongingness to the eight groups also in midlife.
Discussion
General conclusions
Puberty is a milestone in girls’ development. How is this universal process related to
girls’ short- and long-term adjustment? Researchers have long shown an interest in the
psychological and social meanings of girls’ pubertal maturation. Early timing of the pubertal
changes has repeatedly been linked with adjustment difficulties in adolescence, while the
adult implications of girls’ pubertal timing are still not clear. However, a thorough
understanding of the short- and long-term implications of girls’ pubertal timing seems to
require taking other aspects of girls’ maturation into consideration along with biological
pubertal timing. Neglecting to do so might conceal important subgroups of early and later
developers whose adjustment situations might be radically different.
Stattin and Magnusson (1990) proposed one explanation for the link between pubertal
timing and girls’ adjustment. It asserts that through opposite-sex relations with older males
and older peers, early-developing girls encounter peer environments with more advanced
social behaviors (i.e., drinking, drug use, and normbreaking behaviors), and they are brought
into leisure-time settings in which these types of behaviors are more typical. Early-developing
girls adjust to these social settings and peers that are more advanced, and start to act and
behave accordingly. As a consequence, problem behavior among early-developing girls is
seen as engaging in socially more mature behavior, for that age, rather than as deviant
behavior. In this chapter, we tested the proposition that early pubertal timing have limited
implications for concurrent and future adjustment unless it is studied in the light of perceived
maturity and early heterosexual relations. Most of the findings reported here supported this
proposition.
When early-developing girls felt more mature than classmates and when they started
engaging in heterosexual relationships early, they had more adjustment problems of various
kinds in adolescence and they had a family, rather than a career orientation in young
adulthood, compared with other girls. In contrast, early developers who did not feel more
mature than classmates and who did not have early heterosexual relationships had similar
social adjustment in mid-adolescence as other girls. In fact, this group had the least
problematic relationships with parents of all girls, while early developers who felt mature and
who had early heterosexual contacts had the most problematic relationship with parents.
There were instances when the groups differed even in midlife. Previous research has
shown that early developers have higher body mass indexes, diet more often, and are less
physically fit compared to other women at age 43. On the other hand, women were reported to
be similar in most other respects, including education attainments (Johansson & Ritzén,
2005). In the present study, where we investigated pubertal timing in the light of early
22
23
PAPE R II
heterosexual relationships and feeling more mature than classmates, we found evidence that
women do differ in terms of their educational levels at age 43. This is yet another example of
how important it is to look at pubertal timing in relation to other aspects of maturity and to
view pubertal development in its developmental context.
This chapter demonstrates the importance of taking girls’ perceptions of how mature
they are into consideration when examining long-term implications of female pubertal timing.
It seems to be the case that it is only when girls develop early, feel more mature in
adolescence, and have early heterosexual relations that they seem to be at a greater risk of
adjustment problems in adolescence, and where we can see a family orientation, rather than a
career orientation, in adulthood. The findings demonstrate that not considering other factors
together with pubertal timing, has limitations. Rather, incorporating factors such as girls’
cognitions and perceptions about puberty and early heterosexual relationships with
information about pubertal timing is more fruitful. This becomes clear when looking at the
different developmental trajectories incorporating these aspects of maturation. It was a typical
developmental path for early developers who felt mature and who had heterosexual
relationships in mid-adolescence to become mothers early. On the other hand, it was an
atypical developmental path for early-maturers who did not feel mature and did not have
heterosexual relationships in mid-adolescence to become mothers early.
Although the findings in general supported the proposition that early pubertal timing
should be considered in combination with girls’ perceptions of maturity, there were some
instances where girls’ perceptions seemed to be less important than in others. An example is
motherhood in young adulthood. The groups of early developers who did or who did not feel
mature did not differ in terms of how many children they had in young adulthood. For
motherhood, the combination of pubertal timing and early heterosexual relationships seemed
to be most important.
The IDA research program has two main advantages for studying effects of girls’
pubertal timing. First, the project comprises many individuals. This allows researchers to
examine subgroups of early and later developers, which the present chapter is an example of.
Second, and even more important, is the longitudinal design. To our knowledge, no other
project has allowed for prospectively examining implications of girls’ pubertal timing as far
as up to age 43; it is a unique feature of the IDA research program. This design allows us to
explore whether girls’ pubertal timing only plays a role in the timing of events, such as when
women have children, or whether it has implications for how many children women will have
throughout their lifespan.
What does the future hold for early and later-developing girls? This study offers answers
to important questions concerning the role of pubertal timing in women’s adjustment
throughout the lifespan. However, many questions remain. There are studies in the literature
indicating that girls’ pubertal timing is linked with depression, self-image, body perception,
and eating disorders in adolescence (e.g., Kaltiala-Heino, Rimpela, Rissanen, & Rantanen,
2001; Kaltiala-Heino et al., 2003; McCabe, Ricciardelli, & Finemore, 2002). In contrast, little
is known about the long-term implications of pubertal timing on women’s internalizing
problems and psychopathology. This is an area that needs further enquires.
Final remarks
For the 14-year-old girls who were quoted in the beginning of this chapter, their own
and their peers’ maturity were of major concern. Interestingly, the girls’ concern seems to be
warranted, in that maturity is important for adjustment in adolescence, and that it continues to
play this role well into adulthood. Taken as a whole, our findings suggest that it is the
constellation of pubertal timing, perceived maturity, and heterosexual relations – not pubertal
timing, perceived maturity, and heterosexual relationships independently – that is predictive
of concurrent and future adjustment for women. If we would have examined these three
aspects independently, then we would have concluded that the implications of the two aspects
of maturity and heterosexual relationships largely had disappeared by young adulthood, and
also that their implications for adolescent adjustment were smaller than our findings actually
show. Moreover, the findings indicate that neglecting to view pubertal timing in relation to
other aspects of maturity will mask that different early developers have different
developmental paths to adulthood and that these are influenced by their perceptions about
their maturation and early heterosexual relationships.
24
References
25
PAPE R II
Alsaker, F. D. (1995). Timing of puberty and reactions to pubertal changes. In M. Rutter
(Ed.), Psychosocial disturbances in young people. Cambridge: Cambridge University
Press.
Bergman, L. R., & El-Khouri, B. M. (2002). Sleipner – A statistical package for patternoriented analyses,vs. 2.1. Stockholm: Department of Psychology, Stockholm
University.
Brooks-Gunn, J., & Graber, J. A. (1994). Puberty as a biological and social event:
Implications for research on pharmacology. Journal of Adolescent Health, 15, 663-671.
Connolly, S. D., Paikoff, C. M., & Buchanan, C. M. (1997). The interplay of biological and
psychological processes in adolescence. In G. R. Adams, R. Montemayor, & T. P.
Gullotta (Eds.), Psychological development during adolescence (pp. 259-299). Sage:
Thousand Oaks.
Deardorff J., Gonzales, N. A., Christopher, F. S., Roosa, M. W., Millsap, R. E., (2005). Early
Puberty and Adolescent Pregnancy: The Influence of Alcohol Use. Pediatrics, 116,
1451-1456.
Dubas, J.S. (2002, September 1). Long-term versus Temporary effects of Pubertal
Development on Internalizing and Externalizing Problems. Oxford, UK, Paper
presented at the Biennial Meeting of the European Association for Research on
Adolescence.
Edgardh, K. (2000). Sexual behaviour and early coitarche in a national sample of 17 year old
Swedish girls. Sexually transmitted infections, 76, 98-102.
Fergusson, D. M., Horwood, L. J., & Lynskey, M. T. (1994). The comorbidities of adolescent
problem behaviors: A latent class model. Journal of Abnormal Child Psychology, 22,
339-354.
Flannery, D. J., Rowe, D. C., & Gulley, B. L. (1993). Impact of pubertal status, timing, and
age on adolescent sexual experience and delinquency. Journal of Adolescent Research,
8, 21-40.
Galambos, N. L., Barker, E. T., & Tilton-Weaver, L. C. (2003). Who gets caught at maturity
gap? A study of pseudomature, immature, and mature adolescents. International
Journal of Behavioral Development, 27, 253-263.
Galambos, N. L., Kolaric, G. C., Sears, H. A., & Maggs, J. L. (1999). Adolescents’ subjective
age: An indicator of perceived maturity. Journal of Research on Adolescence, 9, 309337.
Galambos, N. L., & Tilton-Weaver, L. C. (2000). Adolescents' psychosocial maturity,
problem behavior, and subjective age: In search of the adultoid. Applied Developmental
Research, 4, 178-192.
Graber, J. A., Petersen, A. C., & Brooks-Gunn, J. (1996). Pubertal processes: Methods,
measures, and models. In J. A.Graber, J. Brooks-Gunn, & A. C. Petersen (Eds.),
Transitions through adolescence (pp. 23-53), New Jersey: Lawrence Erlbaum
Associates.
Graber, J. A., Lewinsohn P. M., Seeley, J. R., & Brooks-Gunn J. (1997). Is psychopathology
associated with the timing of pubertal development? Journal of the American Academy
of Child and Adolescent Psychiatry 36, 1768–1776.
Graber, J. A., Seeley, J. R., Brooks-Gunn, J., & Lewinsohn, P. M. (2004). Is Pubertal Timing
Associated With Psychopathology in Young Adulthood? Journal of the American
Academy of Child and Adolescent Psychiatry, 43, 718-726.
Haynie, D. L. (2003). Contexts of Risk? Explaining the Link Between Girls' Pubertal
Development and Their Delinquency Involvement. Social Forces, 82, 355-397.
Herman-Giddens, M. E., Slora, E. J., Wasserman, R. C., Bourdony, C. J., Bhapkar, M. V.,
Koch, G. G., & Hasemeier, C. M. (1997). Secondary sexual characteristics and menses
in young girls seen in office practice: A study from the pediatric research in office
settings network. Pediatrics, 99, 505-512.
Hockaday, C., Crase, S. J., Shelley, M. C. II, & Stockdale, D. F. (2000). A prospective study
of adolescent pregnancy. Journal of Adolescence, 23, 423-438.
Johansson, T., & Ritzén, E. M. (2005). Very long term follow-up of girls with early and late
menarche. Endocrine Development, 8, 126-136.
Kaltiala-Heino, R., Rimpela, M., Rissanen, A., & Rantanen, P. (2001). Early puberty and
early sexual activity are associated with bulimic-type eating pathology in middle
adolescence. Journal of Adolescent Health, 28, 346-352.
Kaltiala-Heino, R., Marttunen, M., Rantanen, P., & Rimpela, M. (2003). Early puberty is
associated with mental health problems in middle adolescence. Social Science &
Medicine, 57, 1055 – 1064.
Lienert, G. A., & zur Oeveste, H. (1985). CFA as a Statistical Tool for Developmental
Research. Educational and Psychological Measurement, 45, 301-307.
Magnusson, D., Dunér, A., & Zetterblom, G. (1975). Adjustment: A longitudinal study. New
York: Wiley.
Manlove, J. (1997). Early motherhood in an intergenerational perspective: The experience of
a British cohort. Journal of Marriage and the Family, 59, 263-279.
Marin, B. V., Coyle, K. K., Gomez, C. A., Carvajal, S. C., & Kirby, D. B. (2000). Older
boyfriends and girlfriends increased risk of sexual initiation in young adolescents.
Journal of Adolescent Health, 27, 409-418.
McCabe, M. P., & Ricciardelli, L. A., & Finemore, J. (2002). The role of puberty, media and
popularity with peers on strategies to increase weight, decrease weight and increase
muscle tone among adolescent boys and girls. Journal of Psychosomatic Research, 52,
145-154.
Moffitt, T. E. (1993). Adolescence – limited and life-course-persistent antisocial behavior. A
developmental taxonomy. Psychological Review, 100, 674-701.
O'Donnell, L., O'Donnell, C. R., & Stueve, A. (2001). Early sexual initiation and subsequent
sex-related risks among urban minority youth: The Reach for Health Study. Family
Planning Perspectives, 33, 368-275.
O’Laughlin, E. M., & Anderson, V. N. (2001). Perceptions of parenthood among young
adults: Implications for career and family planning. The American Journal of Family
Therapy, 29, 95-108.
Orr, D. P., Beiter, M., & Ingersoll, G. (1991). Premature sexual activity as an indicator of
psychosocial risk. Pediatrics, 87, 141-147.
Pitkanen, T., Lyyra, A-L., & Pulkkinen, L. (2005). Age of onset of drinking and the use of
alcohol in adulthood: a follow-up study from age 8-42 for females and males. Addiction,
100, 652-661.
Statistics on age at first child and marriage available at the Statistics Sweden at www.scb.se.
Swedish. April, 2006.
Stattin, H., & Magnusson, D. (1990). Pubertal maturation in female development. Hillsdale,
NJ: Erlbaum.
Steinberg, L. (1987). Impact of puberty on family relations: Effects of pubertal status and
pubertal timing. Developmental Psychology, 23, 451-460.
Tanner, J. M. (1978). Foetus into man. Cambridge, MA: Harvard University press.
Tilton-Weaver, L. C., Galambos, N. L., & Vitunski, E. T. (2001). Five images of maturity in
adolescence: What does “grown up” mean? Journal of Adolescence, 24, 143158.Tschann et al., 1994
26
27
PAPE R II
Traggiai, C., & Stanhope, R. (2003). Disorders of pubertal development. Best practice &
Research Clinical Obstetrics & Gynaecology, 17, 41-56.
Tschann, J. M., Adler, N. E., Irwin, C. E., Millstein, S. G., Turner, R. A., and Kegeles, S. M.
(1994). Initiation of substance use in early adolescence: The roles of pubertal timing and
emotional distress. Health Psychology, 13, 326-333.
Woodward, L. J. Fergusson, D. M., and Horwood, L. J. (2001). Risk factors and life processes
associated with teenage pregnancy: Results from a prospective study from birth to 21
years. Journal of Marriage and the Family, 63, 1170-1184.
Wu, Z., & MacNeill, L. (2002). Education, work, and childbearing after age 30. Journal of
Comparative Family Studies, 33, 191-213.
PAPER III
115
Klunserna Study
115
05-10-25, 10.22
PAPER III
PAPER III
PAPER III
PAPER III
PAPER III
PAPER IV
115
Klunserna Study
115
05-10-25, 10.23
1
Does appetite affect the association between female pubertal
timing and adult weight status?
Therése Skooga
Håkan Stattina
a
PAPER IV
Center for Developmental Research
Örebro University
S-701 82 Örebro
Sweden
[email protected]
[email protected]
Corresponding author:
Therése Skoog
E-mail: [email protected]
Not for publication:
Phone: +46 19 301091
Fax: +46 303484
1
2
Abstract
Early puberty is associated with subsequent overweight. Few, however, have
investigated factors that could influence this link. The aim of this study was to examine
how appetite and childhood BMI might affect the impact of pubertal timing on adult
weight status. Ninety Swedish females were followed from birth to midlife. Data on
pubertal timing (age at menarche) and BMI (ages 8, 25, and 37 years) were collected
during physical examinations, and parental reports were used to obtain information on
appetites (ages 8 to 16 years). Correlation analyses revealed that body mass index
(BMI) tracks over time. Pubertal timing was linked to BMI in childhood and
adulthood, and also to having a relatively large appetite in adolescence. Hierarchical
regression analyses showed that: (1) pubertal timing was connected to adult BMI over
and above childhood BMI, and (2) appetite moderated the link between pubertal timing
and adult BMI, in that it was only among females with large appetites in early
adolescence that pubertal timing was connected to adult adiposity. Our findings
suggest that pubertal timing is connected to adult weight status irrespective of child
weight status, but only among women with large appetites. This might have
implications for the prevention of overweight.
Key words: puberty, BMI, adiposity, appetite, prospective longitudinal design
2
3
Introduction
3
PAPER IV
Obesity is a significant contributor to ill-health in the world today (Kopelman, 2000).
An important step in preventing obesity and the problems to which it gives rise is to
identify people at increased risk for overweight. In this study, we explore the link
between female pubertal timing (i.e., at what age females experience puberty relative to
their peers) and adult weight status and the roles of appetite and childhood weight
status in this link.
Puberty has been identified as a factor in the development of overweight (Dietz,
1997). When girls enter puberty they gain weight, in the form of lean body mass and
adipose tissue. Early maturing girls seem to gain more weight than their peers. In
adolescence, they are stouter than others (Adair & Gordon-Larsen, 2001; Bini, Celi,
Berioli, Bacosi, Stella, Giglio, et al, 2000; Biro, McMahon, Striegel-Moore, Crawford,
Obarzanek, Morrison, et al, 2001; Bratberg, Nilsen, Holmen, & Vatten, 2007;
Johansson & Ritzén, 2005; Villa, Yngve, Poortvliet, Grjibovski, Liiv, Sjöström, et al,
2007), and this difference persists into adulthood (Johansson & Ritzén, 2005;
Freedman, Khan, Serdula, Dietz, Srinivasan, & Berenson, 2005; Garn, LaVelle,
Rosenberg, & Hawthorne, 1986; Hulanicka, Lipowicz, Kozie, & Kowalisko, 2007;
Must, Naumova, Phillips, Blum, Dawson-Hughes, & Rand, 2005; Wang, Zhao, Liu,
Recker, & Deng, 2006). Is the link between maturing early and having higher adult
body mass index (BMI) merely a result of childhood adiposity?
Because BMI tracks from childhood to adulthood, the link between early pubertal
timing and adult weight status might be due to early maturers already having higher
BMI in childhood. High BMI as early as at age 3 has been linked to early puberty
(Adair & Gordon-Larsen, 2001; Davison, Susman, & Birch, 2003; He & Karlberg,
2001; Juul, Teilmann, Scheike, Hertel, Holm, Laursen, et al, 2006; Kaplowitz, Slora,
Wasserman, Pedlow, & Herman-Giddens, 2001; Lee, Appugliese, Kaciroti, Corwyn,
Bradley, & Lumeng, 2007). Further, most of the research (Campbell, Katzmarzyk,
Malina, Rao, Pérusse, & Bouchard, 2001; Freedman, Khan, Serdula, Dietz, Srinivasan,
& Berenson, 2005; Sandhu, Ben-Shlomo, Cole, Holly, & Davey Smith, 2006), but not
all (Casey, Dwyer, Coleman, Valadian, 1992), shows that BMI tracks over the lifespan
(rs = .40 - .60). If the link between early pubertal timing and adult BMI reflects the
tracking of BMI, then it should disappear when statistically controlling for childhood
BMI. Early research showed a link between pubertal timing and adult BMI, and
assumed that the reasons go beyond early fat mass (Garn et al., 1986). More recent
findings are conflicting. Some studies (Freedman et al., 2003; Must et al., 2005) have
controlled for early adiposity and found no, or only a weak, link, whereas other has
suggested that there is such a link (Kindblom, Lorentzon, Norjavaara, Lönn, Brandberg,
Angelhed, et al, 2006). Given that the latter study was on males, the controversy
indicates that the issue has not been resolved, and that more research is needed.
If pubertal timing is connected to adult BMI irrespective of childhood BMI, a
critical question is whether this depends on other conditions. In a follow-up study of
Norwegians, only early maturers with high waist circumference in early adolescence
had elevated levels of BMI in late adolescence (Bratberg et al., 2007). However,
because waist circumference was measured sometime between the ages 12 and 16,
puberty and waist circumference were confounded in that study. Nevertheless, it shows
that not all early maturing girls have high postpubertal BMI, and consequently that
there might be moderating factors involved in determining which early maturers will
have high postpubertal BMI and which will not.
A person’s BMI depends to a large extent on energy intake and expenditure.
Energy intake is influenced by many factors, including appetite. Appetite, or desire for
4
food, is a complex process controlled by several physiological and psychological
factors (Bray, 2000), and is related to energy intake and BMI (Drapeau, King,
Hetherington, Doucet, Blundell, & Tremblay, 2007; Lee & Song, 2007). In fact,
researchers (Lee & Song, 2007) recently argued that parent-reported appetite can be
used as a surrogate measure of children’s energy intake. Controlling for baseline BMI,
parent-reported appetite predicted the BMI of early adolescents 2 years later, which
suggests that appetite is an independent predictor of future BMI. Thus, if early pubertal
timing is a risk factor for high BMI, there are reasons to believe that appetite may
influence the strength of this, such that early maturing females will have higher BMI if
they have large appetites. To our knowledge, no prior studies have examined
associations between pubertal timing and appetite, or investigated whether pubertal
timing interacts with appetite concerning its association adult weight status.
In this study, we examined how childhood weight status and appetite affect the
association between female pubertal timing and adult weight status using data from a
birth-to-midlife longitudinal study. We posed four specific questions: Does childhood
BMI predict adult BMI? Is pubertal timing associated with BMI in childhood and
adulthood? Is pubertal timing associated with adult BMI, after controlling for
childhood BMI? If early pubertal timing is associated with high adult BMI irrespective
of childhood BMI, does appetite strengthen the link?
Methods
Participants
Participants were from a birth-to-midlife longitudinal study conducted at the Clinic for
the Study of Children’s Development and Health at the Karolinska Hospital and
Stockholm University, Sweden. Participants were recruited by asking every fourth
mother who registered at the Solna Antenatal Clinic between April 1955 and March
1958 to take part in a longitudinal study. Solna is a suburban area of Stockholm.
Almost all the mothers (97%) agreed to join the study. A pilot group of 29 children
was added to the initial participants. The original sample comprised 90 female and 122
male children. At 16 years of age, 84 % were still in the study. Participants are
representative of children in Swedish urban districts in terms of socioeconomic status,
parents’ age, civil status of mother, sibling order, and gestational age and weight
(Karlberg, Klackenberg, Klackenberg-Larsson, Lichtenstein, Stensson, & Svennberg,
1968). Data were collected within four weeks, before or after, the participants’
birthdays.
There was some attrition in the sample. From among the original 90 females, there
was information about height and weight at age 8 and pubertal development for 80
subjects. Of these, information on height and weight was obtained for 68 (85%) and 77
(96%) at ages 25 and 37, respectively. Reasons for not participating were change of
residence, death, and lack of interest. There were no differences between females who
dropped out of the study and those who had participated at ages 25 and 37 with respect
to age at menarche, t78 = .14, p > .10 and t78 = .51, p > .10, respectively, or to weight at
age 8, t78 = .59, p > .10 and t78 = .38, p > .10, respectively.
Measures
Age at menarche. From age 10, girls were asked whether their menstrual
bleeding had started. The question was posed every third month. Menarcheal age is
one of the most common ways of assessing the timing of puberty in girls (Dorn, Dahl,
Woodward, & Biro, 2006). It should be noted, however, that menarche typically occurs
4
5
Table 1 Descriptive characteristics of the study measures
Measure
M
Range
(SD)
13.1
10.7 – 16.1
1.09
Age 8
15.9
12.7-21.3
1.77
Age 25
22.4
16.9-38.7
3.98
Age 37
24.0
18.7-46.6
4.22
3.3
1-4
.83
Pubertal timing
Age at menarche
BMI
Appetite prior to menarche
5
PAPER IV
relatively late in the pubertal process. Hence, it is not a measure of pubertal onset (Dorn
et al., 2006).
Body mass index (BMI). Information on height (in centimeters) and weight
(in kilograms with accuracy of 0.1 kilograms) was collected during medical
examinations (Karlberg, Taranger, Engström, Karlberg, Landström, Lichtenstein, et al.,
1976) at ages 8, 25, and 37 years. Weight was assessed with an accuracy of .1
kilograms on a standing-beam scale. Height was measured with stadiometer. To avoid
age effects, the data was collected within four weeks before or after participants’
birthdays. We used this information to calculate participants’ BMI. The formula for
calculating BMI is weight (in kilograms) divided by height (in meters squared). BMI
correlates strongly with fat mass and percentage of body fat in children, adolescents
and adults (Daniels, Khoury, & Morrison, 1997; Pietrobelli, Faith, Allison, Gallagher,
Chiumello, & Heymsfield, 1998). Consequently, BMI is thought to be a reliable
indicator of body fatness.
Appetite. Overweight adolescents have a strong tendency to underreport their
food consumption (Bandini, Schoeller, Cyr, & Dietz, 1990). Therefore, we used
parent-reported information on girls’ appetites. The original purpose of collecting
information on appetite was the focus on normal growth and related psychological
aspects, such as eating habits and eating problems, which physicians, teachers, and
parents take into consideration (Karlberg et al., 1968). At each year between ages 8
and 16, parents were asked: “Has she a good appetite?” Responses were given on a 5point scale. Response options were: (0) “never”, (1) “seldom”, (2) “sometimes”, (3)
“often”, and (4) “always.” We created a score for appetite one year before menarche
(irrespective of when menarche occurred). The timing of the questioning constituted an
attempt to compare girls’ appetites when they were at approximately the same level of
physical maturity level, thereby avoiding confounding between the puberty and appetite
variables. The average year-to-year correlation for this measure was r = .59, with rs
ranging from .45 to .73.
6
Data analyses
Descriptive characteristics of the study measures are presented in Table 1. The mean
age at menarche (13.1 years) was somewhat below the recent Scandinavian figure (13.4
years; Juul et al., 2006), and somewhat above the recent North American figure (12.8
years) (Herman-Giddens, Slora, Wasserman, Bourdony, Bhapkar, Koch, et al., 1997).
According to its kurtosis and skewness, the scale was normally distributed.
We used logarithmic (ln) transformations of BMI scores at ages 8, 25, and 37 to fit
model assumptions in linear regression analyses (Bratberg et al., 2007). We used
Pearson correlations to examine associations between the study measures. Two
hierarchical regression analyses were performed – one to examine the relative
contributions of childhood BMI and pubertal timing in predicting adult BMI, the other
to test whether appetite is a moderator of the link between pubertal timing and adult
BMI. For the latter test, we created an interaction term between age at menarche and
appetite one year before menarche. The variables were centered before the analysis to
avoid multicollinearity. There was partial drop-out, which made the samples in all the
analyses smaller than the total number of participants.
Some individuals were missing one or more measures included in the analyses. To
ensure the exclusion of these individuals did not affect the pattern of significant
findings, we also performed the analyses in Mplus using FIML (Muthén & Muthén,
2006). This procedure is thought to provide less biased estimates than listwise or
pairwise deletion (Shafer & Graham, 2002). An identical pattern of findings emerged
in these analyses, suggesting missingness is not systematically related to the results
presented here.
Ethical considerations
The Swedish Data Inspection approved the study.
Results
Does childhood weight status predict adult weight status?
Table 2 presents Pearson correlations between the study measures. The analyses
revealed that BMI at age 8 predicted BMI at both ages 25 and 37. The correlations
were moderate. The association between BMI at age 25 and age 37 was strong. Thus,
childhood weight status predicts adult weight status, and weight status at age 25
predicts weight status at age 37 [Table 2].
Is pubertal timing associated with weight status in childhood
and adulthood?
Correlational analyses revealed that age at menarche was associated with BMI at all
ages. Thus, girls with more body fat entered puberty sooner, and the earlier girls
experienced menarche the higher was their BMI in adulthood [Table 2].
6
7
Table 2. Zero-order correlations among all study measures.
BMI age 25
BMI age 37
Age at menarche
Appetite one year
before age at menarche
BMI
age 8
.66***
66
.52***
70
-.23*
78
.20
BMI age
25
.79***
65
-.31*
68
.32**
BMI
age 37
-.33**
70
.27*
72
64
66
† = .06 * p < .05 ** p < .01 *** p < .001
Age at
menarche
-.22†
73
To answer this question, we ran hierarchical regression analyses with BMI at ages 25
and 37 as dependent variables. We entered BMI at age 8 in the first step. It explained
44% and 25% of the variance in BMI at ages 25 and 27, respectively. At the following
step, we entered age at menarche. Together, childhood BMI and age at menarche
explained 47% and 31% of the variance in BMI at ages 25 and 37, respectively. Both
variables emerged as significant predictors of adult BMI. BMI at age 8 emerged as the
strongest predictor. Thus, age at menarche was linked to adult weight status
irrespective of childhood weight status [Table 3].
Table 3. Hierarchical regression predicting log-transformed BMI at ages 25 and 37
from BMI at age 8 and age at menarche.
______________________________________________________________________
Variables in the Model
Standardized Beta
t
Sig
______________________________________________________________________
Predicting BMI at age 25
Step 1
BMI age 8
.66
7.04
< .001
Step 2 Add age at menarche
BMI age 8
.63
6.68
< .001
Age at menarche
-.18
-1.89
.06
Predicting BMI at age 37
Step 1
BMI age 8
.50
4.71
< .001
Step 2 Add age at menarche
BMI age 8
.44
4.12
< .001
Age at menarche
-.24
-2.27
.03
______________________________________________________________________
Predicting BMI at age 25
Model F (1,64) = 49.51, p < .001; R2 = .44 Model F (2,63) = 27.55, p < .001; R2 = .47
Predicting BMI at age 37
Model F (1,66) = 22.19, p < .001; R2 = .25 Model F (2,65) = 14.37, p < .001; R2 = .31
7
PAPER IV
Does pubertal timing have a unique predictive impact on adult
weight status, after controlling for childhood weight status?
8
Is appetite a moderator of the link between pubertal timing and
adult weight status?
First, we examined whether appetite was linked to pubertal timing and concurrent BMI.
Age at menarche was related to appetite (p = .058), in that one year before menarche
there was a trend for early maturing girls to have larger appetites than late maturing
girls. Moreover, appetite at age 8 and BMI at age 8 were not significantly correlated.
Hierarchical regressions predicting BMI at ages 25 and 37 are presented in Table 4. To
test whether appetite is a moderator of the link between pubertal timing and adult BMI,
we entered age at menarche and appetite one year before age at menarche as a first step.
Together, age at menarche and appetite explained 18% and 15% of the variance in BMI
at ages 25 and 37, respectively. Age at menarche emerged as the strongest predictor.
At the second step, the interaction term (age at menarche*appetite) was entered.
Together, age at menarche, appetite, and the interaction term explained 26% and 24%
of the variance in BMI, respectively, and all were significant predictors of BMI at ages
25 and 37. Thus, appetite in early adolescence was a moderator of the link between
pubertal timing and adult weight status [Table 4].
Table 4. Hierarchical regression predicting log-transformed BMI at ages 25 and 37 from
age at menarche, appetite one year before menarche and the interaction.
______________________________________________________________________
Variables in the Model
Standardized Beta
t
Sig
______________________________________________________________________
Predicting BMI at age 25
Step 1 Main effects of age at menarche and adolescent appetite
Age at menarche
-.28
-2.33
.02
Adolescent appetite
.27
2.25
.03
Step 2 Add the interaction
Age at menarche
-.26
-2.33
.02
Adolescent appetite
.35
2.95
.004
Menarche * appetite
-.31
-2.67
.01
Predicting BMI at age 37
Step 1 Main effects of age at menarche and adolescent appetite
Age at menarche
-.29
-2.44
.02
Adolescent appetite
.21
1.78
.08
Step 2 Add the interaction
Age at menarche
-.27
-2.37
.02
Adolescent appetite
.29
2.48
.02
Menarche * appetite
-.31
-2.76
.008
______________________________________________________________________
Predicting BMI at age 25
Model F (2,61) = 6.55, p < .01; R2 = .18 Model F (3,60) = 7.19, p < .001; R2 = .26
Predicting BMI at age 37
Model F (2,63) = 5.60, p < .01; R2 = .15 Model F (3,62) = 6.67, p = .001; R2 = .24
8
9
The interaction between age at menarche and appetite in early adolescence as a
predictor of BMI at age 37 is illustrated in Figure 1. The picture is similar for the
prediction of BMI at age 25. Age at menarche was not related to adult BMI for girls
with small appetites, but was for those with large appetites. Among girls with large
appetites, early age at menarche was linked to high adult BMI, but late age at menarche
was not. Thus, appetite moderated the link between early pubertal timing and high
adult BMI. Apparently, early maturing females only have high BMI if they have large
appetites [Figure 1].
1
0,8
0,6
0,2
Low appetite
0
-0,2
Low age at menarche
High age at menarche
High appetite
-0,4
-0,6
-0,8
-1
Figure 1.
Illustration of the interaction of age at menarche and adolescent appetite for predicting
log-transformed BMI at age 37.
Discussion
Despite growing awareness that early puberty could constitute a risk for future
overweight (Dietz, 1997; Bini et al., 2000; Biro et al., 2001; Bratberg et al., 2007;
Johansson & Ritzén, 2005; Villa et al., 2007; Freedman et al., 2003; Garn et al., 1986;
Hulanicka et al., 2007; Must et al., 2005; Wang et al., 2006), few studies have
investigated factors that might influence the link. In this study, we examined the roles
of childhood weight status and appetite in the relation between female pubertal timing
and adult weight status on the basis of data from a birth-to-midlife longitudinal study.
We found that childhood weight status predicted adult weight status, and that early
pubertal timing was associated with greater BMI in childhood and adulthood, and also
with having a larger appetite in early adolescence. Moreover, pubertal timing was
associated with adult weight status irrespective of childhood weight status and, as
hypothesized, appetite was a moderator of the link between pubertal timing and adult
weight status, in that it was only among females with large appetites in early
9
PAPER IV
BMI at age 37
0,4
10
adolescence, that pubertal timing was connected to adult weight status. Thus, a
thorough understanding of the implications of female pubertal timing for subsequent
weight status seems to require consideration of other factors that act alongside
maturational timing. Neglecting to do so might conceal the possibility that the link
between pubertal timing and adult BMI depends on other conditions, in particular
appetite, and that early maturing females’ adult weight status might be radically
different because of them.
A number of key issues have been addressed in this study. To our knowledge, it is
the first to have linked female pubertal timing and appetite, and it is also unique in
examining the moderating role of appetite in the link between pubertal timing and adult
BMI. Despite growing evidence that pubertal timing and BMI are associated (Adair &
Gordon-Larsen, 2001; Bini et al., 2000; Biro et al., 2001; Bratberg et al., 2007; Daniels
et al., 1997; Davison et al. 2003; Freedman et al., 2003; Garn et al., 1986; He &
Karlberg, 2001; Hulanicka et al., 2007; Johansson & Ritzén, 2005; Juul et al., 2006;
Kaplowitz et al., 2001; Lee et al., 2007; Must et al., 2005; Villa et al., 2007; Wang et
al., 2006), there is limited information on whether girls’ eating behavior influences this
link, although energy intake and expenditure are known key determinants of a person’s
BMI. In this study, we show that appetite seems to be a moderator of the link between
female pubertal timing and adult BMI, in that it is only when early maturers have large
appetites that they have higher adult BMI scores. The BMIs of early maturers whose
appetites were small were found to be similar to those of late maturing girls.
An issue that has been debated for some time is whether or not the link between
maturing early and having higher adult BMI merely reflects early maturers’ higher
prepubertal BMI and the tracking of BMI (Garn et al., 1986; Must et al., 2005). In this
study, childhood BMI was found to be a moderately strong predictor of adult BMI. The
association is similar to that presented for other female samples (Campbell et al., 2001;
Freedman et al., 2005; Sandhu et al., 2006). Two recent longitudinal studies have
examined the role of female pubertal timing for adult BMI, after controlling for
childhood BMI (Freedman et al., 2003; Must et al., 2005). Neither found a substantial
association between pubertal timing and adult BMI irrespective of childhood BMI. By
contrast, but in line with others (Dietz, 1997; Garn et al., 1986; Kindblom et al., 2006),
childhood BMI was not found to fully explain the link between early pubertal timing
and adult BMI in this study. In the case of BMI at age 37, age at menarche
independently explained 6% of the variance. Taking childhood BMI into account, the
association was not as strong, but it was still significant. Thus, the findings indicate
that the link between maturing early and having more adult weight status is not merely
a reflection of early adiposity.
The finding that adolescent appetite influences the link between pubertal timing and
adult BMI is in line with previous research results indicating that adolescent appetite is
an independent predictor of future BMI (Lee & Song, 2007). The mechanisms
underlying this finding are unknown. The answers might lie in hormonal, lifestyle and
psychological factors. Leptin, a protein hormone that is synthesized and secreted by fat
cells, plays a key role in regulating energy intake and expenditure, and increases in
level with body fat and pubertal maturation (Shalitin & Philip, 2003). This is one
example of a hormone that might shed some light on the associations found in this
study. Physical activity is one example of a lifestyle factor that might help explain the
finding. Energy intake and expenditure largely determine a person’s BMI. Early
maturing girls are less physically active than later maturing girls (Baker, Birch, Trost,
& Davison, 2007), and it might be that, given equal energy intake, their energy
balances are more likely to be in surplus than those of late maturing girls (Garn et al.,
10
11
11
PAPER IV
1986). Further, genetic disposition is an additional factor that might explain the links
between pubertal timing, appetite and adult BMI (Parent, Teilmann, Juul, Skakkebaek,
Toppari, & Bourguignon, 2003).
This study has limitations. First, the sample size was limited, which reflects the
very nature of the study – involving long-term follow-up and extensive medical
examinations. It is possible that some of the nonsignificant associations are due to low
statistical power due to the small sample size. Second, the sample is somewhat old.
Because eating habits may have changed since the data on appetite were collected, the
extent to which the findings can be generalized to girls growing up today is unclear.
This problem, however, is inherent in a design with very long-term follow-ups, and is
impossible to avoid. Third, we only used one item to measure appetite and we do not
know how the participants’ mothers interpreted the appetite question. There is a
possibility that mothers reported that their daughter’s appetite was good simply because
their daughter was stout. There is, however, evidence against this. Researchers (Lee &
Song, 2007) have argued that asking mothers about their child’s appetite at the ages we
did is a good proxy for energy intake, and in our study the appetite measure at age 8
and BMI at age 8 were not significantly correlated. Despite these limitations, the
current study makes several advances. Key strengths are: use of a prospective,
longitudinal design; objective measuring of height and weight; and, timing of the
measurement of childhood BMI, which was over 2.5 years before the first girl
experienced menarche.
The present findings raise some questions. Perhaps the most interesting is why late
maturers’ appetites do not seem to affect their future BMI. Why are late maturers
seemingly protected against developing a high BMI? Future research is needed to
understand the relevance of appetite to the link between pubertal timing and BMI; early
puberty is a risk factor, for example, for breast cancer and cardiovascular disease
(Grumbach & Styne, 2003). Perhaps, these associations only hold for, or are stronger
among, a subgroup of early maturers with certain eating behaviors, because these
disorders are linked to overweight and obesity (Kopelman, 2000). These questions, and
also the need to replicate the current findings using a different sample, call for future
research.
To conclude, in this study, we examined associations between childhood and adult
BMI, pubertal timing, and appetite among females. The main finding was that pubertal
timing was associated with adult weight status irrespective of childhood weight status.
Whereas the findings support the idea that early pubertal timing is a risk factor for high
BMI, the also indicated that early maturing females will only have high adult BMI if
they have large appetites. Further, weight status was found to track over time; pubertal
timing was linked to childhood and adult BMI; and, the link between pubertal timing
and adult BMI did not simply reflect early maturing girls’ higher BMI in childhood.
Thus, early puberty seems to have varying implications for adult weight status among
different groups of females. Furthermore, it appears that pubertal timing plays a role in
adult weight status but that the underlying mechanisms in this association and how
pubertal timing interacts with eating behaviors, including appetite, are questions for
future studies. Further understanding of the complex physiological and psychological
processes underlying puberty and appetite is an important next step in trying to find a
way to prevent high BMI among early maturing females.
12
Acknowledgements
We acknowledge with thanks, Dr. William Burk for statistical help, and Professor
Margaret Kerr and Professor Martin Ritzén for valuable comments. Source of funding
was the Swedish Council for working life and social research.
12
13
References
13
PAPER IV
Adair, L. S., Gordon-Larsen, P. (2001). Maturational timing and overweight prevalence
in US adolescent girls. American Journal of Public Health, 91, 642–644.
Bandini, L. G., Schoeller, D. A., Cyr, H. N., & Dietz, W. H. (1990). Validity of
reported energy intake in obese and nonobese adolescents. American Journal of
Clinical Nutrition, 52, 421-425.
Bini, V., Celi, F, Berioli, M. G., Bacosi, M. L., Stella, P., Giglio, P., et al. (2000). Body
mass index in children and adolescents according to age and pubertal stage.
European Journal Clinical Nutrition, 54, 214–218.
Biro, F. M., McMahon, R. P., Striegel-Moore, R., Crawford, P. B., Obarzanek, E.,
Morrison, et al. (2001). Impact of timing of pubertal maturation on growth in
black and white female adolescents: The national heart, lung, and blood institute
growth and health study. Journal of Pediatrics, 138, 636-643.
Bratberg, G. H., Nilsen, T. I. L., Holmen, T. L., Vatten, L. J. (2007). Early sexual
maturation, central adiposity and subsequent overweight in late adolescence. A
four-year low-up of 1605 adolescent Norwegian boys and girls: the Young
HUNT study. BMC Public Health, 7, 54.
Bray, G. A. (2000). Afferent signals regulating food intake. Proceedings of the
Nutrition Society, 59, 373-384.
Campbell, P. T., Katzmarzyk, P. T., Malina, R. M., Rao, D. C., Pérusse, L., &
Bouchard, C. (2001). Stability of Adiposity Phenotypes from Childhood and
Adolescence into Young Adulthood with Contribution of Parental Measures.
Obesity Research, 9, 394-400.
Casey, V. A., Dwyer, J. T., Coleman, K. A., & Valadian, I. (1992). Body Mass Index
from Childhood to Middle Age: A 50-y Follow-Up. American Journal of Clinical
Nutrition, 56, 14-18.
Daniels, S. R., Khoury, P. R., & Morrison, J. A. (1997). The utility of body mass index
as a measure of body fatness in children and adolescents: differences by race and
gender. Pediatrics, 99, 804–807.
Davison, K. K., Susman, E. J., & Birch, L. L. (2003). Percent body fat at age 5 predicts
earlier pubertal development among girls at age 9. Pediatrics, 111, 815-821.
Dietz, W. H. (1997). Periods of risk in childhood for the development of adult
obesity—what do we need to learn? Journal of Nutrition, 127, 1884S–1886S.
Dorn, L. D., Dahl, R. E., Woodward, H. R., & Biro, F. (2006). Defining the boundaries
of early adolescence: A user's guide to assessing pubertal status and pubertal
timing in research with adolescents. Applied Developmental Science, 10, 30-56.
Drapeau, V., King, N., Hetherington, M., Doucet, E., Blundell, J., & Tremblay, A.
(2007). Appetite sensations and satiety quotient: Predictors of energy intake and
weight loss. Appetite, 48, 159-166.
Freedman, D. S., Khan, L. K., Serdula, M. K., Dietz, W. H., Srinivasan, S. R., &
Berenson, G. S. (2003). The relation of menarcheal age to obesity in childhood
and adulthood: the Bogalusa Heart Study. BMC Pediatrics, 3, 3.
Freedman, D. S., Khan, L. K., Serdula, M. K., Dietz, W. H., Srinivasan, S. R., &
Berenson, G. S. (2005). The relation of menarcheal age to obesity in childhood
and adulthood: the Bogalusa Heart Study. Pediatrics, 115, 22-27.
Garn, S. M., LaVelle, M., Rosenberg, K. R., & Hawthorne, V. M. (1986). Maturational
timing as a factor in female fatness and obesity. American Journal of Clinical
Nutrition, 43, 879-883.
14
Grumbach M. M., & Styne, D.M. (2003). Puberty ontogeny, neuroendocrinology,
physiology, and disorders. In Wilson, J. D. & Foster, P. W. (Eds.). Williams
Textbook of Endocrinology. (pp. 1139–1231). Philadelphia: W.B. Saunders.
Herman-Giddens, M. E., Slora, E. J., Wasserman, R. C., Bourdony, C. J., Bhapkar, M.
V., Koch, G. G., et al. (1997). Secondary sexual characteristics and menses in
young girls seen in office practice: A study from the pediatric research in office
settings network. Pediatrics, 99, 505-512.
He, Q., & Karlberg, J. (2001). BMI in childhood and its association with height gain,
timing of puberty and final height. Pediatric Research, 49, 244-251.
Hulanicka, B., Lipowicz, A., Kozie, S., & Kowalisko, A. (2007). Relationship between
early puberty and the risk of hypertension/overweight at age 50: evidence for a
modified Barker hypothesis among Polish youth. Economics and Human Biology,
5, 48-60.
Johansson, T., & Ritzén, E. M. (2005). Very long term follow-up of girls with early and
late menarche. Endocrine Development, 8, 126-136.
Juul, A., Teilmann, G., Scheike, T., Hertel, N. T., Holm, K., Laursen, E. M., et al.
(2006). Pubertal development in Danish children: comparison of recent European
and US data. International Journal of Andrology, 29, 247-255.
Kaplowitz, P. B., Slora, E. J., Wasserman, R. C., Pedlow, S. E., & Herman-Giddens, M.
E. (2001). Earlier onset of puberty in girls: relation to increased body mass index
and race. Pediatrics, 108, 347-353.
Karlberg, P., Klackenberg, G., Klackenberg-Larsson, I., Lichtenstein, H., Stensson, J.,
& Svennberg, I. (1968). Introduction: Design and aims of the study: Description
of sample. Acta Paediatrica Scandinavia Suppl, 187, 9-26.
Karlberg, P., Taranger, J., Engström, I., Karlberg, J., Landström, T., Lichtenstein, H., et
al. (1976). Physical growth from birth to 16 years and longitudinal outcome of the
study during the same age period. In J. Taranger, The somatic development of
children in a Swedish urban community. (pp. 7-76). Doctoral thesis, Universities
of Gothenburg and Karolinska Hospital.
Kindblom, J. M., Lorentzon, M., Norjavaara, E., Lönn, L., Brandberg, J., Angelhed, J.E., et al. (2006). Pubertal timing is an independent predictor of central adiposity
in young adult males. Diabetes, 55, 3047-3052.
Kopelman, P. G. (2000). Obesity as a medical problem. Nature, 404, 635–643.
Lee, J. M., Appugliese, D., Kaciroti, N., Corwyn, R. F., Bradley, R. H., & Lumeng, J.
C. (2007). Weight status in young girls and the onset of puberty. Pediatrics, 119,
e624-630.
Lee, K., & Song, Y-M. (2007) Parent-reported appetite of a child and the child’s weight
status over a 2-year period in Korean children. Journal of the American Dietetic
Association, 107, 678-680.
Muthén, L. K, & Muthén, B. O. (2006). Mplus users guide. Los Angeles, CA, Muthén
& Muthén.
Must, A., Naumova, E. N., Phillips, S. M., Blum, M., Dawson-Hughes, B., & Rand, M.
W. (2005). Childhood overweight and maturational timing in the development of
adult overweight and fatness: the Newton girls study and its follow-up.
Pediatrics, 116, 620-627.
Parent, A-S., Teilmann, G., Juul, A., Skakkebaek, N. E., Toppari, J., & Bourguignon,
J.-P. (2003). The timing of normal puberty and the age limits of sexual precocity:
variations around the world, secular trends, and changes after migration
Endocrine Reviews, 24, 668-693.
14
15
Pietrobelli, A., Faith, M. S., Allison, D. B., Gallagher, D., Chiumello, G., &
Heymsfield, S. B. (1998). Body mass index as a measure of adiposity among
children and adolescents: a validation study. Journal of Pediatrics, 132, 204-210.
Sandhu, J., Ben-Shlomo, Y., Cole, T. J., Holly, J., & Davey Smith, G. (2006). The
impact of childhood body mass index on timing of puberty, adult stature and
obesity: a follow-up study based on adolescent anthropometry recorded at Christ's
Hospital (1936–1964). International Journal of Obesity, 30, 14–22.
Shafer, J. L., & Graham, J. W. (2002). Missing data: Our view of the state of the art.
Psychological Methods, 7, 147 – 177.
Shalitin, A., & Phillip, M. (2003). Role of obesity and leptin in the pubertal process and
pubertal growth – a review. International Journal of Obesity, 27, 869-874.
Villa, I., Yngve, A., Poortvliet, E., Grjibovski, A., Liiv, K., Sjöström, M., et al. (2007).
Dietary intake among under-, normal, and overweight 9- and 15-year-old
Estonian and Swedish schoolchildren. Public Health Nutrition, 10, 311-322.
Wang, W., Zhao, L.-J., Liu, Y.-Z., Recker, R. R., & Deng, H.W. (2006). Genetic and
environmental correlations between obesity phenotypes and age at menarche.
International Journal of Obesity, 30, 1595-1600.
PAPER IV
15
Publications in series Örebro Studies in Psychology
1. Andershed, Henrik, Antisocial Behavior in Adolescence – The Role of Individual
Characteristics. 2002.
2. Trost, Kari, A new look at parenting during adolescence: Reciprocal interactions
in everyday life. 2002.
3. Jensen, Eva, (Mis)understanding and Learning of Feedback Relations in a Simple
Dynamic System. 2004.
4. Wester Herber, Misse, Talking to me? – Risk Communication to a diverse
Public. 2004.
5. Boersma, Katja, Fear and avoidance in the development of a persistent
musculoskele tal pain problem. Implications for secondary prevention. 2005.
6. Jansson, Markus, Insomnia: Psychological Mechanisms and Early Intervention.
A Cognitive-Behavioral Perspective. 2005.
7. Lindblom, Karin, Utbrändhet i normalbefolkningen. Arbets- och individfaktorers
relation i utvecklingen mot eller återhämtning från utbrändhet. 2006.
(Vetenskaplig uppsats)
8. Almqvist, Lena, Children’s Health and Developmental Delay: Positive
Functioning in Every-day Life. 2006.
9. Johansson, Peter, Understanding psychopathy through the study of long-term
violent offenders. 2006.
10. Persson, Andreas, Leisure in Adolescence: Youth’s activity choices and why they
are linked to problems for some and not others. 2006.
11. Ojala, Maria, Hope and worry: Exploring young people’s values, emotions, and
behavior regarding global environmental problems. 2007.
12. larsson, Mats, Human Iris Characteristics as Biomarkers for Personality. 2007.
13. pakalniskiene, Vilmante, Harsh or Inept Parenting, Youth Characteristics and
Later Adjustment. 2008.
14. skoog, therése, On the developmental significance of female pubertal timing.
2008.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement