Programmable Thermostats That Go Berserk? Taking a

Programmable Thermostats That Go Berserk? Taking a
Programmable Thermostats that Go Berserk?
Taking a Social Perspective on Space Heating in Wisconsin
Monica .J Nevius, Energy Center of Wisconsin and Departments ofSociology and Rural
Sociology, University of Wisconsin-Madison
Scott Pigg, Energy Center of Wisconsin
ABSTRACT
This paper describes the results related to space heating and thermostat use from a
study of owner-occupied, single-family residential housing in Wisconsin conducted by the
Energy Center of Wisconsin in 1999. We find that the average self-reported winter
thermostat setting does not vary substantially by type of thermostat used, is predictive of
heating energy intensity in a way that is consistent with expectations, and appears to be a
good indicator of actual thermostat-setting behavior, assuming that the goal is to compare the
behavior of households rather than to gather an accurate report of their actual thermostat
settings. We also find that attitude toward energy conservation appears to have an indirect
effect on household heating energy intensity by way of thermostat-setting behavior. The
results offer further evidence in support of including social and behavioral variables in
traditional engineering-based studies.
Introduction
For more than a decade, scholars have been calling for an increased focus on the
social variables in studies of energy use. Numerous researchers, including Stern and Oskamp
(1987), Rosa, Machlis and Keating (1988), and Shove, Lutzenhiser, Guy, Hackett, and
Wilhite (1998), have urged those who study energy consumption to pay more attention to the
social and behavioral aspects of energy use. Taking such an approach, they point out, can
help us to understand such mysteries as why identical dwellings occupied by
demographically similar households can vary in their energy use by 200 to 300 percent
(Hackett and Lutzenhiser, 1991) or what makes people choose to have an energy audit or
retrofit their houses (Stern and Aronson 1984).
One area of residential energy consumption that could benefit from the insights
provided by including social variables in the analysis is the understanding of behavior related
to thermostat setting. For example, savings estimates for installing a programmable
thermostat in a residential dwelling typically assume that the household members learn how
to program their thermostat, choose to practice temperature setbacks, and are not already
manually setting back theirthermostat. But are these assumptions justified? And if not, is the
installation of programmable thermostats, especially when funded in whole or part with
public money, justified? If public money is to be spent to reduce energy consumption at the
thermostat, on which households should it be spent, and how should they be identified?
This paper focuses on answering these and other important questions related to space
heating behavior using data from the Residential Characterization study, a study of owner-
Consumer Behavior and Non-Energy Effects - 8.233
occupied, single-family residential housing in Wisconsin conducted by the Energy Center of
Wisconsin (ECW) in 1999. The design of this study has been directly influenced by the calls
for more attention to be paid to the social and behavioral aspects of energy use. For each
randomly selected household, ECW conducted a physical audit (tied to a home energy rating)
on the dwelling and gathered utility billing data. The survey included a written questionnaire
given to household members with questions about behavior, beliefs, and attitudes related to
energy use, as well as about individual characteristics such as household income and the age
and education. We also followed up with in-person structured interviews conducted with a
randomly selected subsample of ten percent of participating households in order to better
understand their energy consumption habits and related beliefs.
Theoretical Background and Development of Hypotheses
Since space heating is the single largest user of energy in Wisconsin households (Pigg
& Nevius 2000), this paper focuses on what can be learned about respondents’ energy
consumption for home heating from the combination of social, psychological, economic, and
physical variables that were gathered as part of this study, with a particular emphasis on
thermostat use. This information sheds new light on the prospects for reducing consumption
of home heating, on the role of programmable thermostats in this reduction, and on the
relationships between thermostat-setting behavior and attitudes toward energy conservation
and efficiency.
Organizations that promote the use of programmable thermostats estimate that
installing a programmable thermostat can save 15 percent or more on an individual
homeowner’s energy bill (e.g., Alliance to Save Energy 2000). These figures imply that, in
aggregate, vast amounts of energy could be saved if only homeowners would install
programmable thermostats. Cross & Judd (1997), however, found that the actual energy
savings from programmable thermostats were less than estimated, due to attrition from the
use of programmable thermostats by homeowners and to the fact that customers frequently
already practiced a manual setback. They suggested revising the estimates of savings from
programmable thermostats downward as a result. We explore both our quantitative and
qualitative data in order to better understand the relationship between the use of various
thermostat types (manual or programmable) and the prospects for residential energy savings.
We also test the hypothesis that programmable thermostat users set their thermostats lower
on average than manual thermostat users.
Studies comparing self-reported and observed behavior have shown that self-reported
behavior is a less reliable measurement than observed behavior (Heberlein 1981; Weigel
1983). However, several studies have documented that self-reported thermostat-setting
behavior can be a reasonably good proxy for actual thermostat use. Kempton & Krabacher
(1987) compared self-reports of thermostat-setting behavior with actual thermostat settings
among seven households. They found that, although informants under-reported their
thermostat settings by an average of 2.2°F,adjusting the reported values upward to make up
for this tendency yielded a result that was reasonably close to actual settings recorded from
the thermostats. Gladhart, Weihi & Krabacher (1988) and Lutz & Wilcox (1990) found
similar relationships between actual and self-reported settings.
8.234
—
While our data cannot determine the degree of over- or under-reporting of thermostatsetting on the part of respondents, it can lend credence to the argument that self-reporte4
thermostat-setting data is a reasonably good indicator of actual behavior, and offer evidence
that this self-reported data is at the very least a reliable means by which to compare
households’ thermostat-setting behaviors. In our questionnaire, we asked homeowners to
provide information about the temperature at which they set their thermostats during various
times of the day and night. Combined with information about the hours per week that
someone was home, we calculated an average winter thermostat setting from the self-reported
data. To determine if there is a relationship between self-reported and actual thermostatsetting behavior, we do some exploratory analysis between this average and measured
heating energy intensity.
Studies of the relationship between general environmental attitudes and behavior have
repeatedly shown weak connections between the two at best (Wicker 1969; Ajzen & Fishbein
1980). However, more recent work indicates that attitude studies of more specific behavior,
such as energy conservation or recycling, are likely to yield results showing a greater
connection between attitudes and behavior than studies with a more general focus (Heberlein
and Black 1976; Vining and Ebreo 1992; Oskamp et al. 1991). With the goal in mind of
designing attitudinal measures that might be helpful in predicting energy conservation
behavior, we incorporated a series of attitudinal constructs into the questionnaire. One of
these was designed to measure attitudes toward energy conservation and efficiency, which we
will refer to as “conservation-orientation.” Nevius (2000) found several relationships
between our measure of conservation-orientation and self-reported thermostat-setting
behavior, including a statistically significant correlation between the attitudinal index and the
average self-reported winter thermostat setting. Respondents who scored higher on the
conservation-orientation index reported lower average winter thermostat settings. In order to
determine if this attitudinal measure can indeed be helpful in predicting actual energy
conservation behavior, we explore the relationship between this measure and actual
thermostat-setting behavior by way of the self-reported winter thermostat setting.
Methodology
Survey Design and Sampling
The Energy Center of Wisconsin conducted a survey of 299 households in Wisconsin
in 1998 and 1999 (Pigg and Nevius 2000). The respondents—all of whom were owneroccupiers of single-family, detached housing units—were randomly selected via a multistage stratified sampling design and recruited by telephone. The study was designed to yield
a sample of households that are representative of the state’s population of families living in
Four questions comprised this index: “I am not interested in making energy-saving improvements to
my home”; “It’s just not worth putting on more clothing inthe winter to try to save a little energy”; “I would
only conserve energy if I could not afford to pay for it”; and “I am not interested in making my home more
energy efficient.” The standardized Cronbach’s alpha for this construct was 0.6616, indicating a high degree of
reliability. The summed index scores were reversed so that respondents who exhibited attitudes more strongly
in favor of energy conservation and efficiency received higher conservation-orientation scores (Nevius 2000).
Consumer Behavior and Non-Energy Effects - 8.235
single-family, owner-occupied units in Wisconsin. New construction (defined as houses built
within the last five years) and low-income households (defined as households earning up to
150 percent of the federal poverty level) were oversampled to ensure that there would be
enough cases for eachto make statistically valid inferences about these populations.
Response Rates and Representativeness
Recruiting households that were available and willing to undergo an energy audit
(which required a household member to be home for two to three hours while the energy
audit was conducted) proved challenging, despite ECW’s offering generous incentives.2
Recruiting was conducted via a CATI system using typical RDD protocol. We were able to
complete the recruitment script with 34 percent of households that were within the scope of
the study, and about one out ofevery three of these households agreed to participate.
While the raw data over-represents low-income households and those living in new
construction, we developed case weights for each observation to make the final study sample
as representative as possible of the overall population. These weights were based on a
combination of the 1990 Census and ECW’s 1999 Appliance Sales Tracking (AST) study
(ECW 1999), a large RDD telephone survey which collected demographic data on a sample
of2,214 households in single-family, owner-occupied units.
When we compared the weighted study sample with the AST data, we found the
study sample to be reasonably representative of the larger population of households living in
single-family, owner-occupied homes in the state, with comparable basic demographic and
individual characteristic data such as age, education, income, etc. We also compared the rates
at which qualified households in two samples reported having and using a programmable
thermostat. In 1999, 36 percent of households in the AST sample reported that they had a
programmable thermostat, a rate only slightly higher than the 33.4 percent of households in
the study sample that reported this. Of those households with programmable thermostats, 72
percent of the AST sample and 82.8 percent of the study sample reported that they had used
theirthermostat’s automatic features. These results suggest that while the study sample is not
biased toward households that have programmable thermostats, it may be slightly biased in
favor of the proportion of these households who use the programmable features. A few other
known biases exist in our data that should be kept in mind when generalizing the study’s
results. The weighted sample appears to somewhat over-represent households with more
family members and householders who have lived at their current address for less than a year.
It also appears to under-represent householders who say they are never at home during the
day on weekdays. In addition, two questions from the four-item scale of conservationorientation were included in the AST study; a comparison of the results suggests that our
sample may be somewhat more favorably inclined toward energy conservation than the AST
sample. However, because the data-gathering methods and the order in which the questions
were asked differed—which can affect the distribution of results from attitudinal questions
(Ajzen & Fishbein 1980)—we do not consider this indicator of bias to be conclusive.
Moreover, since the final Residential Characterization study provides representation from
households across almost the entire range of the conservation-orientation scale, we can
2
8.236
Incentives were either $50 or $100, depending on which subgroup the respondent fell into.
—
compare characteristics of those who are less inclined toward conservation-orientation
against those who are more inclined.
—
Results
Thermostat Type and Self-Reported Thermostat-Setting Behavior
Table 1 shows the distribution of types of thermostats in the sample.3 Two-thirds of
households in the sample use manual thermostats to regulate their heating systems, and onethird use programmable thermostats. From questions posed as part of ECW’s AST surveys in
1995, 1997, and 1999, it appears that the saturation of programmable thermostats is rising in
Wisconsin at an average rate of about 2.5 percent per year, from 25 percent in 1995 to 30
percent in 1997 to 36 percent in 1999.
Table 1. Thermostat Distribution and Use
Nt
Type of thermostat used
299
Manual
Programmable
(percent)
66.6
33.4
Percent of those with programmable thermostats
who report using its automatic features
99
Sleeping hours
When someone is awake at home
When no one was home during the day
281
287
268
n/a
(degrees F)
66.9
68.7
66.4
When the household was away on vacation
Self-reported thermostat setpoint (weighted for
weekday hours house is occupied; excluding
198
61.2
60.9
vacation settings)
249
67.8
Mean hours someone is at home weekdays between
8 a.m. and 5p.m.
261
28.9
t Two households without thermostats are not included in the percentages reported on this table.
67.7
Mean reported temperature during...
*
p<.
82.8
65.7
69.4
65.8
*
*
28.1
05
We asked respondents to tell us at what temperatures they kept their home during
sleeping hours, when someone was at home during the day, and when no one was at home
during the day. This information was used to calculate both day and night setbacks and an
average self-reported winter thermostat setting (weighted for the hours respondents told us
someone was at home during weekdays between 8 a.m. and 5 p.m.), to which we will refer as
the “self-reported thermostat setpoint.” The self-reported thermostat setpoint ranged from
59°F to 74°F, with a mean of 68°F. While both households with manual and with
The thermostat types reported in the text and tables were verified by the auditors when they collected
the physical data.
Consumer Behavior and Non-Energy Effects - 8.237
programmable thermostats reported setting back their thermostats at night and during the day
when no one is home, we found that their setback practices varied. Table 2 shows that
households with programmable thermostats are much less likely to keep their thermostats at a
constant temperature, and report steeper setbacks both at night and during the day when no
one is home, than do households with manual thermostats. At the same time, households with
programmable thermostats report slightly higher settings during the day when someone is
home, which we estimate to be the case well over half the time on average. This appears to
largely offset the setbacks during other periods of the day, so that the self-reported setpoint is
nearly the same between the two groups. If these self-reported data are accurate (an issue we
will examine later), it implies that the mere presence of a programmable thermostat in a home
has a minimal effect on heating energy use on average.
Table 2. Crosstabulation of Day and Night Setback Practices by Thermostat Type
Manual
Night setback practices (n=278):*
Setup
Programmable
(percent)
5.9
1.1
No change
Setback 1-4 degrees
Setback 5+ degrees
49,7
22.2
17.2
47.3
Column totals
100.0
100.0
Day setback practices (n=269):*
Setup
No change
Setback 1-4 degrees
Setback 5+ degrees
*
0.5
1.2
53.0
20.8
26.7
39.5
Column totals
100.0
Chi-square tests for both cross-tabulations are significant at p<.OO1.
100.0
Conservation-Orientation and Self-Reported Thermostat-Setting Behavior
Earlier in this paper we mentioned that Nevius (2000) had found relationships
between the conservation-orientation index and several measures of self-reported thermostat
use. The mean conservation-orientation index score of respondents with programmable
thermostats, 14.4 (on a scale from 4 to 16) is slightly higher than that of respondents with
manual thermostats, 13.8; this difference is statistically significant (p<.05). Table 3 shows
correlations among the index and various measures of thermostat setting behavior. These
relationships are mixed, with more conservation-oriented households practicing steeper
setbacks during the day when no one is home but not at night. However, as was mentioned
earlier, there is a negative and statistically significant correlation between conservationorientation and the self-reported thermostat setpoint, indicating that those respondents who
score higher on the pro-energy conservation orientation index report keeping their
thermostats set at lower average temperatures.
8.238
—
Table 3. Correlations Among Conservation-Orientation and Selected Dependent
Variables4
Conservation- Self-reported
orientation
thermostat
setpoint
Conservation-orientation
Self-reported thermostat
setpoint (OF)
Night setback (F°)
Day setback when no one
is home (F°)
Pearson
Correlation
N
Pearson
Correlation
N
Pearson
Correlation
N
Pearson
Correlation
N
*
Continuous
Continuous
measure of night measure of day
setback
setback when no
one is home
1.0000
273
-0.2002
234
*
249
0.0452
258
0.1789
251
1.0000
*
-0.3648
249
**
1.0000
280
-0.3463
**
0.6335
249
260
**
1.0000
267
p<.OO5, ** p<.000S
Thermostat Type, Behavior, Conservation-Orientation, and Actual Energy Use:
Regression Models
While the self-reported data on thermostat settings suggest little difference in the
average winter thermostat setpoint between households with manual or programmable
thermostats, it is possible that these data are not accurate, or that setbacks reported by
households with programmable thermostats occur more regularly than those reported by
households with manual thermostats.
To explore these issues from another angle, we looked at actual heating energy use as
a function of thermostat type, reported thermostat settings, and homeowner attitudes about
conserving energy. Our analysis is restricted to 147 gas-heated homes for which we had good
ability to isolate heating usage from monthly gas usage data.
We fit several regression models to the data, using heating energy intensity
(Btu/fl2/heating degree day) as the dependent variable in order to remove the confounding
effects of house size and climate,
We also included two independent variables in the models to help control for
differences across houses in insulation levels and air leakage: (1) an insulation control
variable, calculated as the total shell conductivity (U-value times area) divided by the house’s
total conditioned area; and (2) an infiltration variable based on a blower door measurement
of air leakage (ft3/minute at 50 Pascals) divided by the house’s total conditioned area. We
would note, however, that because these are observational data, there may be other
unobserved factors that affect heating and may affect the results.
4Adapted from Nevius (2000).
Consumer Behavior and Non-Energy Effects - 8.239
Model 1 (Table 4) shows the results of regressing heating energy intensity on a binary
variable for the presence of a programmable thermostat, The magnitude ofthe programmable
thermostat coefficient suggests that homes with programmable thermostats use about 2.5
percent less energy for heating than homes with manual thermostats (0.197 divided by the
average heating energy intensity of 6.98 Btu/ft2/heating degree day), but the result is not
statistically significant. The 90-percent confidence interval associated with this estimate
suggests that homes with programmable thermostats have somewhere between ten percent
lower and five percent higher heating energy intensity than homes with manual thermostats.
In contrast, there is a strong relationship between our calculated self-reported
thermostat setpoint and heating energy intensity (Model 2). This model suggests that after
controlling for differences across houses in insulation and infiltration, for every degree
change in the self-reported thermostat setpoint the homeowner can expect to save about 0.24
Btu/ft2/heating degree day. This means that an average homeowner from our sample can
expect to save a little over 3 percent from his or her heating bill with every degree of
reduction in their average winter thermostat setting. A widely used rule of thumb is that each
degree of reduction in the thermostat setpoint results in a 3 percent reduction in heating
energy use (DOE 1980)—nearly the same as the percent reduction we calculated for our
sample. Thus, for the subgroup of houses in our sample that are heated with natural gas, the
self-reported thermostat-setting data appears to be a good indicator of actual behavior,
assuming that the goal is to compare the thermostat-setting behaviors of households rather
than to gather an accurate report of their actual thermostat settings.
Model 3 shows that conservation-orientation on its own does not have any
statistically significant predictive power with respect to heating energy intensity. Given that
the self-reported thermostat setpoint is significantly correlated with household heating energy
intensity, Model 4 suggests that, via the mechanism of respondents’ thermostat-setting
behavior, conservation-orientation may nonetheless have an effect on heating energy
intensity. In future research, we will explore a more fully developed model via a path
analytic framework to determine the extent and direction of the indirect effects of
conservation-orientation on physical measures of heating energy intensity.
Table 4. Unstandardized Regression Coefficients for Regression of Heating Energy
Intensity on Selected Independent Variables
Dependent Variable: Heating Energy Intensity (B.t.u./square footIHDD)
Model 1
Model 2
Model 3
Model 4
Coefficients
Coefficients
Coefficients
(t-values)
(t-values)
(t-values)
Coefficients
(t-values)
Independent Variables
Thermostat type -0.197(-0.681)
Self-reported thermostat setpoint
Conservation-orientation
*
8.240
p<.05,
**
p<.OOS
0.180 (2.755)
---
*
Insulation
Infiltration
6.911 (5.950)
1.242 (5.263)
Intercept
2
AdjustedR
2.783
0.510
**
**
0.226 (3.206) *~
--
-0.047 (-0.669)
--
6.507 (5.011) **
1.236 (4.874) **
-9.316
0.516
7.592 (5.641)
**
0.032 (0.452)
6.462 (4,872)
1.196 (4.376)
**
1.262 (4.799) **
3.114
0.491
-12.953
0.532
**
Thermostat Type: Insights from the Structured Interviews
The structured interviews are useful for better understanding thermostat use and the
prospects for energy savings from the installation of programmable thermostats. It was clear
from the interviews that even after going through the home energy audit and filling out the
questionnaire (which described the different kinds of thermostats about which we asked), at
least four of the thirty cases interviewed did not know what a programmable thermostat was,
and others had not been aware that such thermostats existed before the study.
It is interesting to examine how interviewees with manual thermostats responded to
the audit recommendations and savings estimates (based on an 8 hour setback of 5°F)for
programmable thermostats. Fourteen cases both had a manual thermostat and were not
interested in switching to the programmable variety. Of the seven households with manual
thermostats that set their thermostats back manually during the winter, six felt that there
would be no point to installing a programmable thermostat for their households, and several
expressed disbelief in the annual savings estimated by HERS. Of the seven households with a
manual thermostat who indicated during the interview that they kept their home temperature
constant, six told us that they would not adopt a setback pattern if they were to acquire a
programmable thermostat, and none was interested in acquiring one. The following are the
reasons given for not wanting a programmable thermostat:
• They did not believe the savings estimate provided by the HERS audit.
• The payback or the increased convenience were not worththe cost,
• Setting the thermostat would be a “hassle.” For example, there would be no point
since there was almost always someone at home during the day, and they were not
interested in setting the temperature back at night; they were “technologically
impaired” and would be unable or unwilling to program the thermostat; or costs
might actually go up because their schedules varied so much that programming
the thermostat might become impractical, and so their temperature setting would
end up being constant whereas currently they were setting back manually.
• Most oftheir heating comes from a wood stove instead of the furnace.
• They had heard of a programmable thermostat that “went berserk” and overheated
a house.
Of the two cases that indicated an interest during the interview in obtaining a programmable
thermostat, one did not expect it to change the household’s temperature setting habits, but
wanted it for convenience, and the other planned to use it to begin setting back the thermostat
at night and so felt it would reap cost and energy savings. Of those households that already
had programmable thermostats, three were positive about it and felt it saved them money,
while two were unhappy with it and had stopped using the thermostat’s programmable
features. One of these latter cases had lost the instructions, and no one in the household was
able to reset the thermostat when the power went out, so they ended up keeping the
temperature constant. The other case had been unhappy with the default settings on the
thermostat, but had been unable to figure out how to change the settings despite having the
instructions, and so used the override function when they wanted to change the temperature
Consumer Behavior and Non-Energy Effects - 8.241
setting. These two incident$ suggest that programmable thermostats may be too complicated
to operate.
The analyses of the self-reported thermostat-setting data and of the interview data
indicate that the impact of programmable thermostats on heating energy use in Wisconsin is
modest at best. The habits, attitudes, and beliefs expressed in the interviews and the lack of
significant difference between the self-reported thermostat setpoints reported by those with
programmable and those with manual thermostats suggest very little difference in heating
energy use for these two groups.
Conclusions
Several important conclusions can be drawn from our analysis of the winter
thermostat use and space heating data from the Residential Characterization study. First,
while we acknowledge that with our self-reported thermostat-setting data it is impossible to
tell how much respondents over- or under-report their winter thermostat settings, the results
of our regression analysis indicate that because self-reported thermostat setting data is a
strong predictor of household heating energy intensity, it is a good indicator of actual
behavior, assuming that the goal is to compare the thermostat-setting behaviors of households
rather than to gather an accurate report of their actual thermostat settings.
Second, despite the emphasis that has been placed on the use of programmable
thermostats to reduce thermostat setpoints and so save heating energy, respondents with
programmable thermostats report thermostat setpoints that are not substantially different
from those of respondents with manual thermostats.
Third, the results of the regression analyses and the significant correlation between
conservation-orientation and the self-reported thermostat setpoint suggest that respondents’
attitudes toward energy conservation and efficiency may affect their heating energy
consumption by way of their thermostat-setting behavior. The path-analytical model to be
developed in future research will help us to determine the extent and direction of this
relationship.
Taken together, the second and third conclusions suggest a hypothesis. That is, the
installation of a programmable thermostat in a household that does not have a favorable
attitude toward energy conservation and efficiency is not likely to result in a reduction of the
average winter thermostat setting, and therefore not likely to save significant heating energy;
at the same time, the installation of a programmable thermostat in a household that does have
a favorable attitude is also not likely to save significant heating energy, because such a
household would be more likely to already keep a low average winter thermostat setting.
Again, the path-analytical model will help us to determine the validity of this hypothesis.
Among the households that participated in the structured interviews, we found that
more households with manual thermostats than without already practiced winter temperature
setbacks, and that those households with manual thermostats that kept their winter setting
constant would not begin to practice a setback if they were to obtain a programmable
thermostat. We also learned about some of the myriad reasons why many of these households
did not want a programmable thermostat. These details and the conclusions above lead us to
8.242
suspect that the aggregate savings that can be expected from the installation of programmable
thermostats in residential housing is probably quite modest.
Finally, these results provide yet more evidence that combining the study of the social
and behavioral aspects of energy use with more traditional engineering-based approaches can
be a highly informative research strategy that is well worth pursuing.
—
References
Alliance to Save Energy. 2000. Programmable Thermostats.
http://www.ase.org/consumer/house/ thermostat.htm. March 20,
Cross, D., and D. Judd. 1997. “Automatic Setback Thermostats: Measure Persistence and
Customer Behavior.” In The Future ofEnergy Markets: Evaluation in a Changing
Environment, Proceedings ofthe 1997 International Energy Program Evaluation
Conference, Chicago, Ill., August 27-29.
[DOE] U.S. Department of Energy. 1980. Residential Conservation Service Auditor Training
Manual. Washington, D.C.: U.S. Department of Energy, Office ofBuilding and
Community Systems.
[ECW] Energy Center of Wisconsin. 1999. “Appliance Sales Tracking Study.” Madison,
Wis.: Energy Center of Wisconsin.
Ajzen, I., and Fishbein, M. 1980. Understanding attitudes and Predicting Social Behavior.
Englewood Cliffs, New Jersey: Prentice-Hall.
Gladhart, P. M., J. S. Weihl, and S. Krabacher. 1988. “Reported Versus Actual Thermostat
Settings: A Management Perspective.” In Proceedings of the 1988 ACEEE Summer
Study on Energy Efficiency in Buildings, 11:15-28. Washington, D.C.: American
Council for an Energy-Efficient Economy.
Hackett, B. and Lutzenhiser, L. 1991. “Social Structures and Economic Conduct: Interpreting
Variations in Household Energy Consumption.” Sociological Forum 6 (3):449-470.
Heberlein, T. A. 1981 “Environmental Attitudes.” Zeitschriftfur Umweltpolitik 2:241-270.
Heberlein, T. A. and Black, J. S. 1976. “Attitudinal Specificity and the Prediction of
Behavior in a Field Setting.” Journal ofPersonality and Social Psychology 33:47479,
Kempton, W. and Krabacher, S. 1987, “Thermostat Management: Intensive interviewing
used to interpret instrumentation data.” In Energy Efficiency: Perspectives on
Individual Behavior, ed. W. Kempton and M. Neiman. Washington, D.C.: American
Council for an Energy-Efficient Economy.
Consumer Behavior and Non-Energy Effects - 8.243
Lutz, J. and B. A. Wilcox. 1990. “Comparison of Self-reported and Measured Thermostat
Behavior in New California Houses.” InProceedings ofthe ACEEE 1990 Summer
Study on Energy Efficiency in Buildings, 2:91-100. Washington, D.C.: American
Council for an Energy-Efficient Economy.
Nevius, M. 2000. Household Energy Use and Attitudes toward Energy Conservation:
Implications for Voluntary Energy Conservation Programs. Paper presented at the
American Sociological Association Annual Conference, Washington, D.C., August
12-16.
Oskamp, S., Harrington, M. J., Edwards, T. C., Sherwood, D. L., Okuda, S. M., and
Swanson, D. 1992. “Factors Influencing Household Recycling Behavior.”
Environment and Behavior 23:494-519.
Pigg, S. and M. Nevius. 2000. “Energy and Housing in Wisconsin: a Study of Single-Family
Owner-Occupied Homes.” Madison, Wis.: Energy Center of Wisconsin
(forthcoming).
Rosa, E. A., Machlis, G. B., and Keating, K. M.. 1988. “Energy and Society.” Annual Review
ofSociology 14:149-72.
Shove, E., Luztenhiser, L., Guy, S., Hackett, B., and Wilhite, H. 1998. “Energy and Social
Systems.” In Human Choice and Climate Change, Vol. II, ed. S. Rayner and E. L.
Malone. Seattle: Battelle Pacific Northwest Laboratory.
Stern, P. C., and Aronson, B. 1984. Energy Use: The human dimension. New York: W. H.
Freeman and Company.
Stern, P. C. and Oskamp, S. 1987. “Managing Scarce Environmental Resources.” In
Handbook ofEnvironmental Psychology, Volume 2, ed. D. Stockal and I. Altman.
New York: Wiley.
Vining, J., and Ebreo, A. 1992. “Predicting Recycling Behavior From Global and Specific
Environmental Attitudes and Changes in Recycling Opportunities.” Journal of
Applied Social Psychology, 22:1580-1607.
Weigel, R. H. 1983. “Environmental Attitudes and the Prediction ofBehavior.” In
Environmental Psychology: Directions and Perspectives, ed. N. R, Feimer and E. S.
Geller. New York: Praeger.
Wicker, A. W. 1969. “Attitudes vs. Actions: The relationship ofverbal and overt behavioral
responses to attitude objects.” Journal ofSocial Issues 25:41,
8.244
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