STATE APPROPRIATIONS: IMPLICATIONS FOR TUITION AND FINANCIAL AID POLICIES By Matthew J. Foraker Copyright © Matthew J. Foraker. 2009 A Dissertation Submitted to the Faculty of THE CENTER FOR THE STUDY OF HIGHER EDUCATION In Partial Fulfillment of the Requirements For the Degree of DOCTOR OF PHILOSOPHY In the Graduate College THE UNIVERSITY OF ARIZONA 2009 2 THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE As members of the Dissertation Committee, we certify that we have read the dissertation prepared by Matthew J. Foraker entitled: State Appropriations: Implications for Tuition and Financial Aid Policies and recommend that it be accepted as fulfilling the dissertation requirement for the Degree of Doctor of Philosophy. _______________________________________________Date:____4-10-2009__ Dr. John J. Cheslock _______________________________________________Date:____4-10-2009_ Dr. Gary Rhoades _______________________________________________Date: ____4-10-2009_ Dr. Cecilia Rios-Aguilar _______________________________________________Date: ____4-10-2009_ Dr. Richard J. Kroc II Final approval and acceptance of this dissertation is contingent upon the candidate’s submission of the final copies of the dissertation to the Graduate College. I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement. ________________________________________________Date:____4-10-2009_ Dissertation Director: Dr. John J. Cheslock 3 STATEMENT BY AUTHOR This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at the University of Arizona and is deposited in the University Library to be made available to borrowers under the rules of the Library. Brief quotations from this dissertation are allowable without special permission provided that accurate acknowledgement of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the copyright holder. SIGNED: Matthew J. Foraker 4 ACKNOWLEDGEMENTS To faculty of the Center for the Study of Higher Education: your commitment to scholarship has been inspiring and motivating. A special thanks to Gary Rhoades for conducting excellent and thought provoking classes that encourage students to ask deeper questions and expand their grasp of the big pictures. Most importantly, thanks to my adviser, John Cheslock, without whose commitment and unwavering support I could not have completed this work. To Errol Anderson, whose willingness to take a stand for my possibilities taught me to believe in my ability to climb over the wall after tossing my hat. Finally, to my parents who taught me to value education and my own development and gave me the confidence to pursue my highest aspirations. 5 TABLE OF CONTENTS LIST OF ILLUSTRATIONS …………………………………………… 6 LIST OF TABLES ……………………………………………………… 7 ABSTRACT …………………………………………………………….. 8 CHAPTER ONE – INTRODUCTION ………………………………... 9 CHAPTER TWO – LITERATURE REVIEW …………….………… 19 Introduction ……………………………………………………. 19 Historical Overview ……………………………………………. 20 The Costs of Higher Education ………………………………… 30 State Appropriations ……………………………………………. 34 Tuition …………………………………………………………… 37 Enrollment Management ………………………………………. 43 Related Studies ………………………………………………….. 47 Pertinent Theoretical Frameworks …………………………….. 50 Conclusion ……………………………………………………….. 53 CHAPTER THREE – DATA AND METHODOLOGY ……………… 55 Introduction ……………………………………………………… 55 Research Questions ……………………………………………… 55 Methodology ……………………………………………………… 57 Data Sources ……………………………………………………… 65 Variables …………………………………………………………. 67 Sample Populations ……………………………………………… 71 CHAPTER FOUR – RESULTS ..…………..…………………..………. 76 Introduction ……………………………………………………… 76 Institutional Level Analysis – Four Year Institutions …….….. 77 Institutional Level Analysis – Two Year Institutions ………… 84 Student Level Analysis – Four Year Institutions .…………….. 93 Student Level Analysis – Two Year Institutions ……………… 101 Summary of Results.……………………………………………. 104 CHAPTER FIVE – CONCLUSION ………………………….…………`107 Introduction……………………………………………………… 107 Key Findings…………………………………………………….. 107 The Big Picture …………………………………………………. 109 Implications ……………….……………………………………. 111 Future Research ………………………………………………… 113 Conclusion ………………………………………………………. 115 APPENDIX A . IPEDS INSTITUTIONAL LEVEL DATA SOURCES….…………………….……………………… 117 APPENDIX TABLES …………………………………………………… 122 REFERENCES ………………………………………………………… 152 6 LIST OF ILLUSTRATIONS Figure 1: Loan and Grant Share of Student Aid…………………………27 7 LIST OF TABLES Table 1. Average Higher Education Tuition and Fees………………………….. Table 2. The IPEDS Institution Sample Populations…………………………… Table 3. The NPSAS Sample Populations: Four-Year Institutions…………..…. Table 4. The NPSAS Sample Populations: Two-Year Institutions ………..….... Table 5. Institutional Level Analysis – Four Year Institution Descriptive Statistics……………………….. …………………… Table 6. Institutional Level Analysis – Four Year Regressions: In-State Tuition State Appropriations Coefficient................ …… Table 7. Institutional Level Analysis – Four Year Regressions: Out-of-State Tuition State Appropriations Coefficient…………. Table 8: Institution Level Analysis – Two Year Institution Descriptive Statistics ………………………. …………………… Table 9. Institutional Level Analysis – Two Year Regressions…… ………........ Table 10: Institution Level Analysis – Fixed Effects Regression ………………. Table 11. Student Level Analysis – Descriptive Statistics: Four-Year Institutions Table 12. Student Level Analysis – Linear Regression: In-State Students at Four-Year Institutions ………………………….………..…….. Table 13. Student Level Analysis – Linear Regression: Out-of-State Students at Four-Year Institutions …………………………..…… Table 14. Student Level Analysis – Linear Regression: In-State Students At Two-Year Institutions …………………………………..….…. Table 15. IPEDS / NPSAS / Fixed Effects Regression Analysis Comparison ……………………………………………. Appendix Table 1. Institution Level Analysis Complete Four Year In-State Regression……………………….. Appendix Table 2. Institution Level Analysis Complete Four Year Out-of-State Regression…..……………… Appendix Table 3. Institution Level Analysis Complete Two Year In-State Regression……….………………. Appendix Table 4. Student Level Analysis Complete Four Year In-State Regression…..…………………… Appendix Table 5. Student Level Analysis Complete Four Year Out-of-State Regression………….……….. Appendix Table 6. Student Level Analysis Complete Two Year In-State Regression……………….……….. Appendix Table 7. Descriptive Statistics – IPEDS Institutions…………………. Appendix Table 8. Descriptive Statistics – NPSAS Students …………………… 39 72 74 75 78 80 83 85 87 89 94 98 99 103 108 118 119 120 121 123 125 127 137 8 ABSTRACT Over the past 30 years the costs of higher education have climbed faster than the rate the inflation. As these costs have risen, state appropriations for public institutions have not kept pace. While not declining in real dollars, as a portion of meeting the expenses of funding public higher education, state appropriations have been steadily falling over the past three decades. Not surprisingly, during this period tuition at public colleges and universities has risen dramatically, leading to concerns about access to higher education, in particular for students of low income backgrounds. The literature contains many studies highlighting the increasing costs and tuition charged by public colleges and universities. Little has been written about the specific relationship between the level of state appropriations at a particular institution and the pricing and financial aid policies it then adopts. By analyzing the data for public institutions in the Integrated Postsecondary Educational Data System (IPEDS) as well as data for specific students in the National Postsecondary Student Aid Survey (NPSAS) for five school years spanning 1989-1990 to 2003-2004, this study conducts a quantitative analysis to create a predictive model capable of forecasting the impact of changes in state appropriation on institution pricing and financial aid policy. In an environment where the continued decline of state appropriations as a portion of meeting educational costs is a real possibility, such forecasting ability may prove invaluable in crafting policies to insure access to higher education for certain student populations. 9 CHAPTER ONE INTRODUCTION In the past few decades, state appropriations to public universities have fallen dramatically as a portion of institutional budgets, from 60 percent in 1969 to 24 percent in 1995 (Hearn, 1998; Heller, 2002). In 2008 dollars, from 1978 to 2008 average in-state tuition at public four-year schools rose from $2225 to $6185 or 279 percent, while instate tuition at two-year public institutions grew from $1040 to $2361 or 227% (US College Board). In response to the public outcry over the escalation of higher education prices at rates dramatically faster than the consumer price index, politicians have convened committees, delivered speeches, and published letters on the importance of the affordability of a college education. Some have introduced legislation to cap price increases as well as implement policies that would penalize institutions that raise tuition above a specified percentage. At the same time, they allocate state appropriation levels that leave a gap in meeting the costs institutions incur. Views differ regarding the causes of the escalating costs of higher education. Whether it follows from the advancement of technology and the need to educate for an increasingly complex knowledge economy workforce, or from schools competing for reputations of prestige and excellence, or from other causes, costs are increasing. As state appropriations fail to keep pace, as we will see, schools do not simply make up the difference with tuition. However, clearly the level of state appropriations received by an institution plays a role in its tuition price, and the trends in financial aid programs affect the ability of students to obtain a degree. 10 Higher education is a service oriented endeavor that requires not only expertise in the subject matter, but individuals capable of providing quality instruction in an environment conducive to learning. Labor intensive interaction between teacher and student is often necessary for effective education. Like health care, which requires human interaction between care giver and patient, the nature of education involves a “cost disease” (Baumol & Blackman, 1989) relative to operations that can be automated or mass produced, such as the manufacture of physical products, food, or other fixed items. Technological advances may have in fact exacerbated the costs of providing higher education. Skilled professionals must be hired to support increasingly sophisticated laboratories, classroom equipment, computer centers, libraries, and other technical support services (Rhoades, 1998). Instead of producing the same result at lower costs, higher education chases a moving target that grows more expensive to produce. Contributing to the increase in the cost of higher education is the competition that exists between institutions of similar type. Many schools, in particular four-year public flagship research universities, operate in an environment of intense competition for prestige and a reputation for educational excellence (Ehrenberg, 2001). In an “arms race” to attract the best students and faculty, schools experience pressure to outspend their rivals to produce more attractive campuses, better facilities, first rate Web sites, superior athletic programs, and other enticements to improve their standing in rankings such as those published by US News and World Report. The pressure to “pursue excellence” among the nation’s colleges and universities may have become so institutionalized that even schools substantially down the prestige hierarchy embark on efforts to “climb the 11 ladder.” In what Bowen called the “revenue theory of cost” in his 1980 book, schools generate as much revenue as possible via fundraising, tuition, bond elections in local communities, various grants and other donations, or other sources. Once this revenue is understood, the full amount is allocated towards a prioritized “wish list” until costs meet the total revenues obtained (Bowen, 1980). According to this theory, only the amount of revenue generated restricts the costs that can be incurred. In this study I implicitly examine the influence of three interrelated broad forces, 1) the rising costs of providing higher education, 2) the state appropriations that are not keeping up with these increasing costs, and 3) the implications this gap has for the pricing (tuition) set by public institutions and the financial aid made available to students in meeting those prices. As we will see, both the setting of a tuition price and the level of state appropriations received by a public university follows a complex set of factors including local politics, state resources, university budgets, and competing interests for public support. The financial aid provided to students is also determined by a complex system that combines Federal grants and loans, state merit and need programs, and institutional aid. The model produced in the study quantifies the relationships between variables representing the factors and seeks to explain some of the influences that exist between them. The relevance of the study will only increase given the scenarios likely to exist in the future. The outlook for state appropriations for higher education suggests a continuation of the trends of the past few decades. In funding public higher education, states face tremendous pressure for scarce resources from other important and competing 12 community needs including K-12 education, state infrastructure, correctional facilities, and Medicaid. States often turn to higher education as a budget balancing tool during recessions and other difficult budget situations. In the face of these stagnating state appropriations, schools must either prevent costs from rising or increase revenues from other sources. As we have seen, costs have not been reduced. A natural alternative revenue source is tuition and fees. However, to simply infer that tuition is replacing state appropriations in meeting the costs of instruction overlooks the complexity of higher education pricing and financial aid. The nature of the influence of state appropriations levels per student at a public institution on its tuition price can be expected to vary based upon operating environment in which it functions. Institutions in different states may face different governance structures linked to legislatures and political relationships with different priorities and agendas. Large research universities with prestigious business schools, medical schools, or particular departments recognized as the best in their field have a broader range of choices in response to a change in funding than, for example, a local community college serving a rather fixed population of nearby students. Tuition has risen dramatically in the past three decades at both private and public higher education institutions. While in total dollars the increases have been the greatest at elite, private schools, in percentage terms the public institutions have raised tuition as much and in many cases more, including tuition at two-year community colleges. At the same time that costs and tuition levels have risen, financial aid policies have shifted at the federal, state, and institutional level. Federal financial aid provided directly to students has shifted dramatically from a grant focused approach to that of loans which must be 13 repaid, and the federal aid system has significantly relaxed its requirements on student need, allowing even relatively affluent students to obtain low cost loans for attending college (Hearn, 1998). States have used tobacco settlements and lotteries to fund state financial aid programs that emphasize student merit over financial need. Both of these trends have impacted access and attendance patterns of lower income students, who are loan averse and disadvantaged against wealthier students in merit based programs. At the institutional level, the use of tuition discounting and institutional grants has grown significantly (Redd, 2000). While partially motivated by the desire to provide aid to students based on need, institutional aid is significantly motivated by enrollment management considerations, and often students are awarded grants based on non-need factors such as academic ability, demographics including gender, and indeed, the student’s ability to pay high tuition (e.g. someone from a wealthy family or an out-ofstate student) (Wilkinson, 2004). In this study, at both the institutional level and the student level, I examine the impact of state appropriation support per full time student provided to public higher education institutions on the list tuition prices they set and the financial aid packages provided to students. At the institutional level, I use data from the Integrated Postsecondary Education Data System (IPEDS) to run regressions quantifying the relationship between state appropriations and tuition prices taking into account the selectivity of the organization and the higher education governance system of the state in which they operate. The study separately examines five years during which the NPSAS studies were conducted, 1989/90, 1992/93. 1995/96, 1999/2000, and 2003/04. For these 14 years, I compare the results with those obtained using student level data from the National Postsecondary Student Aid Surveys (NPSAS) for students attending the same set of institutions. The study will clarify our understanding of the role state appropriations plays in pricing and financial aid policy. At the level of institutions, the analysis quantifies the impact of state appropriations on 1) the listed “sticker price” tuition at public universities and community colleges while controlling for the selectivity of the institution and the governance structure of the state in which it operates. In-state tuition is distinguished from out-ofstate tuition, and four-year schools are run separately from two-year schools. For the four-year institutions, I also disaggregate the schools by classification of selectivity to discern any differences between the different types of schools, anticipating that more selective institutions may respond differently when changes to their appropriations occurs. At the level of the individual students, the analysis quantifies the impact of state appropriations on 1) the listed tuition price, 2 ) the “net tuition price the student paid after receiving a grant (not to be repaid) from the school, 3) the institutional grant provided to the student, and 4) the “unmet need” as reported by the data set, equal to the difference between the student’s expected family contribution and the price the student actually incurred. The NPSAS data set with student level data allows the analysis to operate at more refined level, permitting one to control for student characteristics including socioeconomic status (using expected family contribution as a proxy), academic merit (using their cumulative grade point average at the school), and ethnicity and gender. 15 Since the student analysis can control for these factors, I anticipate student level data to paint a different picture and will explore why these differences may exist. Purpose of the Study The purpose of the study is to examine the relationship between state appropriations and higher education pricing and financial aid policies. Using detailed quantitative analysis of prices charged and aid awarded to students attending public colleges and universities across the country, the study will produce a model with the ability to generate forecasts of the probable pricing and financial aid consequences that will occur at a public two-year or four-year college or university when it faces changes in state appropriation levels. This tool will provide additional clarity in the understanding of the relationships between state appropriations, governance, and the financial transactions that occur between students and the institutions they attend. I investigate the theories applicable to the analyses and address how the theories might explain the findings of the model. Finally, the study uses existing price and aid response literature to explore the implications of the findings on access and participation of different student populations in higher education. Organization of the Study The study consists of five chapters. Chapter one provides an introduction to the study including its purpose, approach, and its contribution to the study of higher education. Chapter two presents a brief overview of the history of higher education in the United States before conducting a review of the pertinent literature related to public support for higher education, its costs, pricing, and the financial aid programs designed to 16 help students enroll. I then present the political, economic, and other organizational theories applicable to the study and relevant to the interpretation of its findings. Chapter three provides the details of the methodology of the study and the data sets utilized. I explain the selection of the sample populations and variables used to produce the econometric models. Chapter four provides a detailed discussion of the results obtained and relates them to the frameworks and theories under consideration. Chapter five provides an overview summarizing the results and the conclusions one can draw from the study along with the implications they may have for policy makers. Summary and Significance of This Study An abundance of literature exists regarding rising costs of providing higher education and the tuition levels charged to students. Many studies have also documented the declining share of state appropriations in public university budgets. Descriptive statistics regarding tuition levels, financial aid awards, state appropriations, and attendance patterns have been published, but little quantitative work has been performed in an effort to create predictive models that can forecast the likely impact of changes in one area (in this case, state appropriation levels) on what occurs in other areas (tuition prices set and financial aid packages provided). Further, the extent to which this relationship may vary for different kinds of students (family income, gender, ethnicity, academic preparation, in-state status) as well as for different kinds of institutions (selectivity, two-year or four-year, state governance environment) has yet to be explored. A deeper understanding of the implications of the growing gap between public support for higher education and the costs of providing it, and by implications I refer to 17 the resulting tuition prices and financial aid packages made available to aspiring students, is critical to providing policy makers the information necessary to produce legislation that leads to intended consequences to the extent possible. Without such knowledge, policies may be set that have unintended and undesirable consequences for access to higher education and higher education programs in general. Public support for higher education remains politically viable, and it is likely to continue to be an important component in assisting the majority of American college students. However, the mechanisms for financing college attendance are clearly shifting in focus and intent. Federal assistance has moved from a predominantly need justified grant orientation towards the use of subsidized (and unsubsidized) federally backed loans with less strict need qualifications (Hearn, 1998). States have moved towards direct aid to students based on merit. As these shifts occur and occur differently in different states across the country, what can be said about the impact of changes in unrestricted state appropriations for higher education, traditionally the greatest source of financing higher education? A quantitative analysis producing a model providing predictive capability of the effect of such changes on pricing and financial aid can lead to the next step in understanding impact on access and participation in higher education in the United States. The implications have substantial relevance in light of the reality that the current downward trend in the share of state appropriations in financing public higher education will continue. Even prior to the financial upheaval in the fall of 2008, states have been contending with tremendous pressure for support from other important community needs 18 including transportation infrastructure, the public kindergarten through high school education system, departments of corrections, and Medicaid for a growing population of seniors. All of these will continue to compete with higher education for scarce public resources, suggesting that the findings of this study will be important and remain important into the foreseeable future. 19 CHAPTER TWO LITERATURE REVIEW Introduction The review begins with an overview of the history of public higher education finance and summarizes the key developments that led to the current system. After introducing the broad historical background of higher education, and I provide a brief overview of the period from the industrial revolution and the need for a more skilled workforce through the world wars, the cold war, and the Higher Education Act of 1965. I then present the history of higher education finance in greater detail, discussing the major federal legislation and the financial aid programs produced at the federal, state, and institutional level. I discuss the more recent trends associated with the costs of higher education and cover the two fundamental cost theories most often referenced when discussing the growing expenses associated with providing higher education. I then address the developments taking place regarding state appropriations for public colleges and universities and how the trends in costs and state appropriations are impacting tuition policies. After I explore the trends and the literature discussing tuition in detail, I discuss the development of the enrollment management function and review the literature that examines the ways in which both institutions and students are responding to all of the above. Finally, I explore the economic and institutional theories pertinent to these issues and provide a conclusion of the review and an overview of its implications for public higher education in the United States. 20 Historical Overview Prior to the Civil War, higher education existed as an option only available to a select few born into wealth and privilege. Most Americans regarded the completion of high school as ample education necessary to enter the workforce in the nineteenth century economy that was heavily agricultural, retail, and trades and crafts to apprentices taught by their employers as their employers were once taught as apprentices a generation earlier. For the small, elite (and male) privileged set of youth destined to become lawyers, doctors, politicians, and other civic leaders, prestigious private schools existed to develop leaders and professionals with a curriculum rooted in philosophy, Latin, culture, critical thinking, and the classics of western civilization. The private schools at this time still struggled for enrollment and admitted most who applied and could pay the fees. As the nation shifted to an industrial economy and the rise of the corporation, organizations emerged that had management and control structures requiring managers, accountants, engineers, scientists, bookkeepers, attorneys, and other professionals necessary to run increasingly sophisticated manufacturing lines, oil refineries, chemical plants, and other operations. These positions required workers educated beyond the high school level. On July 2, 1862, President Abraham Lincoln signed The Morrill Land Grant Act of 1862 which gave the states federal lands for the establishment of colleges that not only provided the traditional academic curricula, but also offered formal training in agriculture, mining, engineering, home economics, and other “practical” subjects (Wilkinson, 2004). Over time all of the states in the nation built land grant universities 21 that subsidized the cost of educating students with state appropriated funds provided to schools as unrestricted support per student enrolled. This funding allowed universities to charge little or no tuition. In some states, such as New York and California, tuition remained free for state residents well into the twentieth century. Enrollment at public universities steadily grew as the industrial economy matured and electricity, the automobile, and the department store became common place. Large numbers of soldiers returning from World War Two combined with Cold War pressure for technological superiority to the Soviet Union dramatically increased the nation’s interest in having as educated a population as possible. On June 22, 1944, President Franklin D. Roosevelt signed the Servicemen’s Readjustment Act of 1994, better known as the GI Bill. The landmark legislation, controversial at the time, fueled unprecedented levels of participation in higher education and altered the perception of higher education itself, now seen as viable for many if not all instead of a select few. The Soviet Union’s launch of Sputnik in 1957 intensified the government’s interest in an educated workforce, particularly in math and science. In 1958 the National Defense Education Act (NDEA) provided aid to education at all levels, but particularly targeted at programs seen as benefiting national security, including technical education, math, science, and also modern foreign languages. The legislation also established the National Defense Student Loan Program (NDSL), which offered low interest loans to qualified students. President Lyndon Johnson’s “War on Poverty” effort noted that college graduates earned more income than high school graduates. The support of widespread participation 22 in higher education fit into the “Great Society” agenda that included Medicaid, the Voting Rights Act, and the Elementary and Secondary Education Act (Hannah, 1996). This combined with the view that an educated population enhanced society led to the profound and far reaching Higher Education Act of 1965 (Hearn, 1998; Wilkinson, 2004). President Johnson signed the legislation in November of 1965, and it funneled billions of dollars into the higher education system, its expansion, and financial support for students to pay the price of attendance. Title IV of the Act established the Federal Guaranteed Student Loan (GSL), the Equal Opportunity Grant (EOG), and the College Work Study Program (CWS) (Hearn, 2002). All of these programs sent participation in college spiraling upward. In 1940, 16 percent of eighteen to twenty-one year olds attended college; by 1970, the portion had tripled to 48 percent enrolled in college (Wilkinson, 2004). The nation’s culture shifted further. Not only was higher education now seen as viable for all, but many came to believe everyone, including ethnic minorities and women, should have the right to obtain a college degree. Starting in the 1970s public sentiment began to shift. The financial aid programs enacted by the Higher Education Act of 1965 provided the families of students the means to finance higher education, lessening the need for universal low tuition (or in some states, no tuition at all). In 1968 free market proponent Milton Friedman published an article suggesting that the laws of supply and demand should apply to higher education, with students paying the full price of the market determined cost of their education (Friedman, 1968). The “High Tuition High Aid” model argued that universal low tuition was inherently inefficient, allowing wealthy students who could easily pay the full cost to 23 enjoy highly subsidized prices. While few endorsed the severity of Friedman’s suggestion, support grew for what came to be called the “rationalization” of tuition, i.e. having tuition prices more closely reflect the costs incurred by the institution, with the belief that removing the “blanket subsidy” for all students led to improved equity and efficiency (Hearn, Griswold, & Marine, 1996). In addition to the voices advocating tuition prices more in line with the costs of the education were sentiments that supported the idea of introducing competition into the market for higher education. The Higher Education Amendments of 1972 radically altered the landscape of higher education finance by attaching aid to the student instead of the institution, based on the argument that empowering students with choice would foster competition and provide incentive for higher education institutions to improve quality and prestige to attract better students (Wilkinson, 2004). The amendments created the Basic Education Opportunity Grant Program (BEOG), which placed the financial support at the disposal of the enrolling student to use at the institution of choice. The legislation also renamed the National Defense Student Loan as the National Direct Student Loan, and the Economic Opportunity Grant was renamed as the Supplemental Education Opportunity Grant (SEOG). In a prescription for trouble, the law allowed proprietary (for-profit) schools to use Title IV financial aid programs. The for-profit institutions became most adept at assisting students in obtaining financial aid. Caught in the middle without the means to pay tuition but failing to qualify for federal need based aid, middle income families successfully exerted political pressure to succeed with The Middle Income Student Assistance Act (MISAA) of 1978. The Act removed income 24 restrictions on federally subsidized student loans (GSL), which sent the use of such loans soaring from $3.389 billion in 1975-76 to $10.623 billion in 1980-81 (Hearn, 1998). In 1981, income restrictions were restored, but not at levels prior to the MISAA legislation (Wilkinson, 2004). Viewing government as “the problem” and the private sector as the solution, the Reagan administration battled Congress regarding higher education funding and financial aid, arguing that the Department of Education was not necessary and calling for its dissolution. Secretary of Education William Bennett asserted that “greedy” higher education institutions raised tuition knowing students would receive federal aid to pay for it, an assertion never proven (McPherson & Shapiro, 1991) for public institutions and refuted for private schools (Wilkinson, 2004). Reagan failed to eliminate financial aid or the Department of Education, but he succeeded in producing a profound shift in the culture regarding higher education. The student, not society, received the bulk of the benefits of a college degree. Accordingly, the students and their families should bear the bulk of the cost (Hearn, 1998). In 1980, Congress wanted to honor a man known for his advocacy of higher education participation and attainment on the part of those requiring financial assistance, Senator Claiborne Pell of Rhode Island. Pell was largely responsible for the creation of the Basic Educational Opportunity Grant (BEOG), so the Higher Education Amendments of 1980 renamed the program as the Pell Grant, a direct financial assistance program available to all students seeking to enroll with demonstrated financial need. The legislation also created the Parent Loan for Undergraduate Students (PLUS) that allowed 25 middle income families to borrow $3,000 a year for each dependent child regardless of the family income (Wolanin, 1998). The Higher Education Amendments of 1992 sought to rationalize and standardize the financial aid system as much as possible. The political environment of 1992, however, lacked the cohesive thrust for education that existed during Johnson’s “Great Society” (Hannah, 1996) and despite its efforts, the results continued the “alphabet soup” of overlapping federal and state efforts to promote college enrollment. The legislation did create a single methodology for the determination of the cost of college attendance and the ability of a dependent or independent student to pay for it. It mandated for all students seeking aid to solely use the Free Application for Federal Student Aid (FAFSA), and it standardized the calculation of student need to a single method called the Federal Need Analysis. The act also provided a system for the pro-ration of Pell Grants and Stafford Loans to adjust for attendance not fitting a complete academic year. Additional name changing took place, with the renaming of the Guaranteed Student Loan Program to the Federal Stafford Loan in honor of Senator Robert Stafford of Vermont. The Stafford Loan Program also created an additional loan program for students who did not qualify for the established, federally subsidized loans. These were called the Stafford “Unsubsidized” loans and while less generous than the subsidized loan programs, they still offered attractive loan options for otherwise ineligible students (Hannah, 1996). In 1997, President Clinton signed the Taxpayer Relief Act which established tax credits for higher education expenses. It created the HOPE Scholarship Tax Credit giving up to a $1500 tax credit for each of the first two years of college attendance. The 26 interest on loans for education was made tax deductible, and it created the ability for a family to establish an educational “individual retirement account” (IRA) for each dependent child under 18. The law also permitted the withdrawal of funds from existing IRA’s without penalty if the funds were used for college attendance. What did all of this mean for the students? Perhaps the single greatest difference experienced by students and their families over the past 30 years was the federal financial aid system’s substantial shift from a need-based grant orientation towards a significantly less need sensitive loan orientation. This change resulted in students graduating with debt burdens so great that James Hearn wrote of the system graduating “indentured servants” (Hearn, 1998). Some middle and lower income students were graduating with four-year degrees and college loans exceeding $40,000, more than their annual salaries upon graduation. The shift in the federal financial aid system from grants to loans shifted the share of total aid provided to students to one based more upon loans than upon grants (Baum & Payea, 2003), as shown in Figure 1. 27 Figure 1. The Percent Share of Total Financial Aid, 1982/83 to 2002/03 Percent Share of Total Aid 65 Percentage 60 55 50 Grants 45 Loans 40 35 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 30 Academic Year (Graph produced with data from Trends in Student Aid, College Board, 2003) As the Federal government shifted to a loan orientation for helping students pay for college, the states, using settlements from tobacco lawsuit settlements and/or revenues generated from state lotteries, shifted from a need-based focus towards a focus on rewarding students based on academic credentials and performance while attending college. In 1993, Georgia created a program called “Helping Outstanding Pupils Educationally” (HOPE) (Heller, 1997). Florida followed quickly with a program called Bright Futures. Other states quickly implemented merit based programs. While state aid based on need did not decline in absolute dollars, state merit aid increased at a dramatically higher rate. From 1982 to 1998, state need based aid increased 88 percent in real dollars. In the same time period, state merit aid increased 336 percent, most of it since 1993 (Heller, 1998). 28 Both the federal aid trend towards loans and the state aid trend towards merit based rewards tend to leave lower income students, who are less likely to qualify for merit aid and are loan averse (Heller, 2003) without any means of financial assistance. Higher education institutions seeking to maintain access as well as maintain diversity in their classrooms resorted to institutional aid to recruit students. Institutional aid (especially grants) is the fastest growing segment of financial aid in higher education (Heller, 1997). Using NPSAS (National Postsecondary Student Aid Study) data for 1989-90 and 1995-96, Heller found that institutional grants to undergraduates climbed from $386.6 million to $928.4 million at public institutions, the fastest growing type of financial aid to students (Heller, 1997). Institutional aid to students exists within the context of enrollment management (which is discussed in depth later in this chapter), which may (or may not) consider student need a priority. While institutions may target aid with financial need as one factor, other factors including the academic merit, ethnicity, gender, or other aspects of the student’s background, (in particular, the student’s ability to pay as an attractive means to boost net tuition revenue) may be considered in determining the aid package. Further, universities do not appear to employ institutional financial aid practices that compensate for shortcomings in state level financial aid policies. Doyle, Delaney and Naughton (2003) found that instead of using institutional aid to fill gaps or replace aid reductions caused by state shifts in financial aid, the schools instead distributed financial assistance in a manner that reinforced the state position, complying with and furthering the state’s philosophy regarding which students should receive what amounts of financial assistance. In those states moving 29 towards merit based aid programs, institutions only exacerbate the shortfalls created for need based aid funding (Doyle, Delaney, & Naughton, 2008). The debate over the financing of higher education has continued to the present day. The stratification of institutions in terms of prestige has continued to grow with the gap widening between private schools and public universities. Economic inequality in the United States increased during the Bush administration, the period where the wealthiest ten percent of Americans realized a 24% increase in income while the median income for the nation fell 8% and income for the lowest ten percent of Americans dropped 17% (Bureau of Labor Statistics, 2008). Despite the efforts to rationalize the financial aid system, in particular the Higher Education Amendment of 1992, to this day we still have a system where college access and choice remains stratified by socioeconomic status and ethnicity (Thomas & Perna, 2004). Hearn (2001) noted that instead of a unified and coherent system of programs with clarity and “distinctiveness,” the system continues to consist of overlapping efforts and objectives. The federal programs implemented during the Great Society Era have been modified, adjusted, and renamed. The states have each implemented their own financial assistance programs with shifting goals and objectives. As a result, we have a confusing array of federal, state, and institutional programs to promote college enrollment (Perna, et. Al; 2008) that can overwhelm graduating high school seniors considering higher education, in particular those who would be the first time college students in their families. 30 The Costs of Higher Education The costs of providing a higher education have risen dramatically in the last few decades. Baumol and Blackman (1989) first coined the term “cost disease” to refer to labor intensive activity that cannot realize the productivity gains made possible by technology or the streamlining of processes that can occur in other endeavors such as manufacturing or information technology. According to the “technology theory of costs,” Baumol and Blackman assert that services fitting this category such as education, health care, legal representation, and others, will continue to experience rising costs above the consumer price index (Baumol and Blackman, 1989). Bowen’s revenue theory of cost offers a different perspective. This view sees higher education institutions seeking to maximize revenue in the pursuit of the greatest degree of excellence possible. The increasing costs of higher education develop from the intense competition for status and prestige, exacerbated by the increasing stratification of higher education institutions (Trow, 1984) and widely publicized and read rankings of colleges and universities in the US News and World Report Magazine (Ehrenberg, 2002). The top private elite universities with massive endowments have the ability to spend extraordinary sums of money to increase status to attract and recruit top talent for both faculty and students. Princeton’s endowment per student in 2006 was estimated at $1.9 million, with several other top private universities also having endowments per student exceeding $1 million (NACUBO Endowment Study, 2006). With respect to the public universities, the revenue theory implies that institutions will already be setting tuition at levels which maximize revenue, so an examination of the affect of state appropriation 31 changes on tuition prices should find little or no relationship. The institutions maximize revenue subject to constraints placed on the student body enrolled in terms of gender, ethnicity, academic merit, or other institutional imperatives (Bowen, 1980). Bowen’s theory has drawn a lot of attention and is plausible given the extraordinary stratification of educational costs in higher education institutions. Schools at different levels in the hierarchy of prestige spend vastly different amounts teaching their students (Winston, 1999). While the students at highly selective private universities may pay far greater levels of tuition, they get what they pay for (Winston, 1999) because the well endowed institution has the resources to incur educational costs far exceeding the amount of the tuition. While a community college may charge $500 for a public (and state subsidized) program that costs $800 per student, an extremely expensive, selective, and elite private school, while charging $40,000 tuition, may be incurring costs exceeding $100,000 for each student who enrolls. As noted earlier, at the top of the hierarchy, private institutions and (increasingly) public flagship research institutions have endowment incomes per student exceeding the listed tuition price (NACUBO Endowment Study, 2006). Bowen’s revenue theory is an appealing perspective in an environment of such competition. Universities have engaged in an academic prestige “arms race” with a scoreboard published each year that Ehrenberg (2002) has shown to influence the number of applications a school receives and the enrollment yields of students admitted. The race pits institutions into competition not only for top students, but top faculty, and while the public universities have all but lost any capability to compete with top private institutions 32 for faculty (Alexander, 2001), they certainly compete with each other. In the competition for faculty, universities can use means other than direct compensation by offering lower teaching loads or other incentives. Zemsky (1994) coined the phrase of “The Academic Ratchet” regarding the downward pressure on faculty teaching loads and its contribution to the escalation of educational costs. While the revenue theory may not fully explain the rising costs of providing higher education, it does predict activities that universities have begun to enhance their revenue streams. Public schools now generate endowments. They conduct alumni fundraising campaigns or other programs to solicit donations. Public research universities seek lucrative research grants and business ventures at technology parks. At the undergraduate level, schools have a greater profit orientation at student recreational centers, student unions and cafeterias, parking fees, and others (Levy, 1995; Slaughter & Rhoades, 2004). The key point to be made here for this study is that the revenue theory, the diversification of revenue streams, and shifting financial aid all suggest that the relationship between state appropriation levels and tuition at a particular institution, while intuitively appealing, is anything but straightforward. Some question the extent to which the revenue theory really applies to higher education institutions. Getz and Siegfried (1991) listed six competing (with the revenue theory as well as each other) explanations for the escalation of costs: 1) the cost disease introduced by Baumol and Blackman, 2) a higher education shift towards more expensive disciplines, 3) shortages of higher education inputs, 4) faculty and administrators in charge having inflated desires for quality, 5) poor management in higher education, 6) 33 costs arising from government regulations creating expanded duties for higher education (Getz & Seigfried,1991). While the desire of the institution for quality may fit into Bowen’s ideas, the rest are distinctly different and more thematically in line with cost disease. In a study designed to compare the explanatory effectiveness of the revenue theory of cost with that of cost disease, Archibald & Feldman (2008) examined unit costs for a large variety of goods and services in the United States from 1949 to 1995. If cost disease were the major factor, other products and services also affected by the theory should experience similar cost increases. If Bowen’s theory were more significant, cost increases should be quite different from those to whom it applied. They found that: Cost per student in higher education follows a time path very similar to the time path of other personal service industries that rely on highly educated labor. This is entirely consistent with the cost disease explanation of the rise in cost in higher education. This explanation is based on strong economy-wide influences that affect industries that tend to experience lagging productivity growth and that rely on highly educated labor, not on characteristics of higher education itself. (Archibald & Feldman, 2008) The findings cast doubt on the universal application of Bowen’s revenue theory to higher education as a whole, but do not go so far as saying that it does not apply in certain groups of institutions or certain departments within an institution. A study that applies the ideas of Archibald and Feldman that also compares the results of different groups of schools after disaggregating the institutions by selectivity, size of endowment, or other measures associated with prestige may yields interesting and useful findings, as would a study that examines the tuition price response and financial aid changes that occur when institutions experience a change in public funding. 34 State Appropriations As mentioned earlier, the Morrill Land Grant Act of 1962 provided federal land (and funds from the sale of federal land) for the purpose of created state supported public higher education rooted in the practical advancement of knowledge if fields of use to the economy. States provide funding appropriations, usually based on the number of students, to offset the cost of the education and permit schools to keep tuition well below the actual cost of the education. Known as universal low tuition, the practice rose from the belief that education benefited society as a whole, furnished the national economy with workers it required, and improved the country’s ability to compete (Heller, 2002). At all non profit higher education institutions, both private and public, educational costs equal price plus a subsidy (Winston, 1999). Historically, to keep the price well below the cost, private schools used endowments and fundraising, and public schools used state appropriations. While this continues to apply to private institutions, during the last three decades the financing of public higher education has shifted away from taxpayer subsidies towards individual students and their families. The conversation has shifted from education as a public good to be financed predominately at public expense, to education as an investment on the part of the students in anticipation of enhanced employment opportunities and income. While Milton Friedman’s desire to see the free market fully implemented in higher education has not occurred, during the last three decades the promotion of competition both on the part of students as consumers and on the part of institutions as suppliers has clearly shifted the production and consumption of higher education in the direction of the free market. 35 State appropriations to public colleges and universities have not kept pace with increasing costs (Selingo, 2003). While in absolute dollars, state subsidies have not fallen, as a share of public institution budgets, state appropriations have declined since 1980 (Hearn, 1998; Heller, 2002 ). In 1969, the state covered 60 percent of the expenditures share, while tuition and fees covered 15 percent, a spread of 45 percent. The 1969 difference between state appropriations and tuition and fees was 45.1%, which peaked at 47.4% in 1978 and has been falling ever since. By 1995, it had fallen to 24.1%, half of its prior value (Heller, 2004). The conservative shift in the view of the federal government’s role in social programs that started with the Reagan administration saw the federal government transfer responsibility for many programs to state or local governments (Weertz & Ronca, 2006). From 1980 to 1992, the higher education expenditures covered by the federal government declined from 18 percent to 14 percent (Hossler, 1997). Partially due to political sentiment, and partially out of necessity, states started to view higher education, now competing with Medicaid, corrections, and other social services for funding, as a more “optional” program in tough times of recession and economic scarcity. Seen as a “balancing wheel” component of the state budget, higher education faced likely reductions during recessions or budget crisis situations. (Selingo, 2003; Callan, 2002). One consequence of this perspective is that in many cases the tuition levels set by state legislatures and higher education institutions hardly follows carefully crafted strategy or rational policies. Rather, as Mary McKeown (1982) noted, higher education public institution pricing policies are often “backed into” after addressing budget short falls, and 36 Griswold & Marine (1996) also found that “many states appear to increase tuition in reaction to situational factors rather than developing a rational policy or even a planned response to perceived needs.” Johnstone (1993) observed numerous obstacles to the implementation of a hightuition high aid model. He noted that aid for middle and low income students may not keep pace with rising tuition, and that the rapidly rising costs of attending college had political repercussions, an assertion confirmed when Congressmen McKeon tried to introduced legislation punishing universities if they raised tuition beyond a certain “affordability index” (McKeon, 2003). Clearly, the tuition levels of public higher education institutions, in particular the in-state tuition paid by the families of voters, are influenced by political considerations. McLendon, Hearn, and Hammond (2008) conducted a panel study of the tuition prices at 162 of the nation’s top public universities over the period from 1984 to 2002. They found that the setting of tuition levels is a complex function of many factors including a state’s population and economic climate, state appropriations and aid policies, higher education governance structures, and importantly, the state’s political environment. Higher levels of minority representation in state legislatures were associated with lower tuition prices at the state’s public flagship institutions (McLendon, Hearn, & Hammond, 2008). Weertz & Ronca (2006) also found political factors played a key role in the tuition prices at public universities. They found that the support for higher education in the form of state appropriations per student was most associated with a) campus commitment to public service and community outreach, b) the higher education governance system, and c) gubernatorial and legislative support. 37 As state appropriations decline in relevance as a share of meeting educational costs, simple budget balancing demands that other revenue sources must increase in relevance, the most obvious replacement being net tuition revenue. At one extreme, where public support fully covers the institution’s budget, we have schools providing universal low tuition to all applicants, a system most closely approached during the 1960s and 1970s with the large expansion of higher education participation. The opposite extreme, Friedman’s desire for a free market in higher education, would have students and their families pay the full cost of their education, minus whatever financial aid is made available to them on a case by case basis. Without getting into the full implications of removing the “public” from public higher education institutions, the political reality of voters with children desiring a college education is likely to keep federal and state support in place for the foreseeable future. The continued slippage in the share of state appropriations as a portion of university budgets, however, remains likely (Selingo, 2003). Tuition In the face of rising costs and the leveling off of state support, the notion that institutions simply pass the difference off to students in the form of higher tuition oversimplifies the complex response of public universities and students to the development. That said, tuition levels have indeed risen at rates far exceeding the consumer price index and the cost of educating the students (Ehrenberg, 2002). The stagnation of state appropriations combined with the cultural shift to the idea that students receive the bulk of the benefits of a higher education has increased scrutiny on 38 universal low tuition. Opponents of universal low tuition noted that wealthy and upper income students, those most likely to participate in higher education, could easily afford to pay higher tuition levels and free the taxpayer by reducing the subsidy higher education receives (Hearn & Longanecker, 1985). Theoretically, raising tuition and also raising aid for those with financial need results in a more precise allocation of assistance to those that required it. Wealthy students would pay an amount equal to (or at least much closer to) the true cost of their education. Hearn, Griswold & Marine (1996) noted that financial distress in state budgets often results in tuition levels and policy being “backed into” as well as regional differences in the state policies. The northeastern and upper Midwestern regions of the country performed greater “rationalization” tuition than southern and southwestern states. The setting of tuition levels, in particular in-state tuition, occurs inside of a political context. As shown in Table 1, more recently the percentage increases in tuition and fees at public four-year institutions exceed those at private four-year schools. How do the students and their families react to such tuition increases? During the environment of heated debate about financing higher education, Hearn & Longanecker (1985) conducted a study of the enrollment affects of pricing policies. Not surprisingly, they found that students responded to tuition prices differently depending on the incomes of their families, with the demand elasticity the greatest (i.e. the sharpest drop in enrollment when tuition is raised) for low income students. Leslie and Brinkman (1987) 39 Table 1 Average Higher Education Tuition & Fees (Enrollment Weighted – Constant 2007 dollars) School Year Private Four-Year % Chg Public 4-Year % Chg Public 2-Year % Chg 1977-78 $9,172 $2,225 $1,040 1982-83 $9,872 8% $2,194 – 1% $1,007 –3% 1987-88 $12,808 30% $2,699 23% $1,343 33% 1992-93 $15,416 20% $3,444 28% $1,647 23% 1997-98 $17,823 16% $4,022 17% $2,026 23% 2002-03 $20,778 17% $4,715 17% $1,926 –5% 2007-08 $23,712 31% $6,185 51% $2,361 41% (Source, US College Board, Trends in Higher Education Series, 2007) reviewed the results of 25 studies between 1967 and 1985 to create an average student response to increases in tuition, finding that on average student demand drops 0.6 % when tuition is increased $100. Ten years later Heller (1997) replicated this finding with results varying between 0.5 and 1.0 percent reductions in enrollment given a $100 rise in tuition. Heller’s study went further and presented five key findings of critical importance to the “high tuition, high aid” policy. First, regarding a student response to an increase in tution, in 1983 dollars, a $100 increase produced a 0.7% drop in enrollment. Second, student response to a change in financial aid, while more complex because it varies on the type of aid involved, was weaker than that of a tuition increase. Importantly, even with a change in aid to neutralize an increase in tuition, keep net price constant, the increase in the sticker price of tuition caused “sticker shock” and a drop in enrollment. As one would expect, loans and work study had an even weaker effect than grants (Heller, 1997), a fact exacerbated by the federal policy of increasing the share of federal financial aid to 40 loans (Hearn, 1998). Third, students from lower income families are more sensitive to price increases. Only at the higher SES levels does sensitivity to price disappear. Poorest students react strongest to grants, even stronger than tuition (likely that they know they will qualify). Fourth, rising tuition in higher education is producing the “trickle down” mentioned earlier, where meritorious students select schools at lower levels of selectivity because of price. Many of the poorest students are giving up on higher education entirely and obtaining unskilled work in the labor market. Naturally these are aggregate trends with exceptions existing at all levels. Ethnicity and race do play a role, but one must be cautious of the interplay between these and income. Family income is the most influential variable (Heller, 1997). Finally, he finds variation across institutional type with significantly stronger student response to price increases at community colleges. Price increases are experienced as a portion of the total cost. A $100 increase at an institution charging $20,000 annual tuition is not the same as a $100 increase at an institution charging $1,000 (Heller, 1997). Summarizing, the model demoralizes and discourages participation of the lower income students. Using the 1995-1996 and 2003-2004 National Postsecondary Student Aid Surveys (NPSAS), Wellman, Desrochers, and Lenihan (2008) found that among dependent students, participation in higher education of those from families earning less $20,000 fell from 14.5% to 12.8%, and the impact is not limited to the lower income families. Over the same period the enrollment of graduating high school seniors from families earning less than $80,000 fell from 66.8% to only 49.1% (Wellman, Desrochers, and Lenihan, 2008). Gerald & Haycock (2007) found that the portion of Pell grant recipients at more 41 prestigious flagship public schools among all Pell grant recipients fell from 0.83 in 1992 to 0.63 in 2003, with even more dramatic declines for students with minority backgrounds (Gerald & Haycock, 2007). Intuition readily anticipates that non-resident (out-of-state) student response would differ from that of in-state students. Zhang (2007) specifically examined nonresident demand at public universities. Conducting analyses at the national, state, and institutional level, he found that at the national level, a one percent increase in nonresident tuition reduced non-resident enrollment by 0.90 percent. However, at the state level no such relationship occurred, and at the institutional level the elasticity was far smaller, where a percent increase in non-resident tuition only reduced non-resident enrollment by 0.20 percent. He also found the interesting result that an increase in the resident (in-state) tuition increased non-resident enrollment, suggesting that in-state tuition may serve as a proxy for prestige, a result supported when he disaggregated the schools by type and selectivity. For the Research I Doctoral universities or highly competitive schools, non-resident student demand exhibited virtually no elasticity with enrollment falling only 0.04 percent for the former and 0.11 percent for the latter when non-resident tuition was increased by one percent. The generation of funding streams other than tuition is relevant to this study. The National Association of State Universities and Land-Grant Colleges, analyzing IPEDS revenue data, compared the total revenue percentages at public four-year institutions between the 1990/91 academic year and the 2000/01 academic year, dividing total revenue into three sources: 1) state appropriations, 2) tuition, and 3) other revenue. State 42 appropriations fell from 36.1% of total revenues to 29.8%. While tuition revenue increased, the rise was significantly smaller than the drop in state appropriations, climbing from 15.3% to 17.7%, an increase of only 2.4% in the face of a 6.3% drop in state appropriations. The “other revenue” stream increased 3.9% from 48.6% to 52.5%. I want to underscore the point that public four year institutions now receive over half of their total revenue from sources other than state appropriations and tuition. Political factors appear to play a role in the setting of in-state tuition prices. Studying panel data for 162 public flagship universities over the period from 1984 to 2002, McLendon, Hearn, and Hammond (2008) found that higher levels of minority representation in state legislatures were associated with lower levels of tuition. They also confirmed the findings of Koshal and Koshal (2000) that increases in state appropriation levels decelerated increases in resident tuition. Regarding out-of-state enrollment and tuition prices, they found that schools enrolling larger shares of out-of-state students charged higher in-state tuition levels, a finding consistent with Zhang’s study on nonresident enrollment demand. Strictly focused on nonresident students, Zhang found an increase of 1% in resident tuition is associated with a 0.88% increase in nonresident enrollment (Zhang, 2007). In particular for out-of-state students, tuition appears to serve as a proxy for institutional quality (Zhang, 2005). As institutions become more prestigious, out-of-state and international student demand rises, and universities charge these students a substantial premium over the in-state tuition level charged to state residents. As state-appropriations fall, the political mission of serving state residents can be expected to have a suppressing effect on using in-state tuition to make up the 43 difference. In the context of the politics of state legislatures governing university tuition prices, one would expect in-state tuition to react more sluggishly than out-of-state tuition prices in efforts to increase revenue. The downward political pressure on the levels of in-state tuition at public universities does not apply to out-of-state tuition for the simple reason that it does not apply to residential students and their voting parents. The distinction between the in-state student and out-of-state student is important to this study. The setting of in-state tuition levels has been shown to be a function of the political environment and the university campus’s engagement in the community. The highly inelastic out-of-state student demand found by Zhang suggests that such students are a ripe source of net tuition revenue provided the institution has the necessary level of prestige and reputation. The balance between in-state tuition and enrollment and out-of-state tuition and enrollment has become important for public universities, and many now have departments dedicated to a relatively new and increasingly important function in higher education, enrollment management. Enrollment Management During the period of rapid expansion of participation in higher education that followed the Higher Education Act of 1965, a natural stratification of institutions occurred as demand far outstripped supply. The prestigious Ivy League and other elite private schools, bombarded with many more applicants than they could admit, became highly selective, and an “axis of prestige” developed with expensive and highly selective private institutions at the top and inexpensive community colleges at the bottom. At the 44 same time that the Higher Education Act of 1972 encouraged the competition of institutions for students now free to apply their financial aid to the school of their choice, competition amongst students for admission into the best college possible also flourished. Barron’s and other publications started publishing books categorizing schools by selectivity, with the measure serving as a universally understood proxy for status, prestige, the quality of the education, and the prospects of the students upon graduation. In 1993, the magazine US News & World Report started publishing detailed rankings of institutions, a publication devoured by high school students and their parents as they reached college age. In a matter of less than twenty years, an institution’s place in the prestige hierarchy began to significantly affect its operating environment in terms of applications received, percentages of admitted students choosing to enroll, and ability to recruit top faculty (Ehrenberg, 2001). The formulas used by the magazine to rank the institutions rewarded selectivity and the costs expended per student with profound implications. Summarizing, the increased stratification of institutions and the introduction of competition on the part of both suppliers and customers complicated the “market” for higher education. The need to manage this growing complexity transformed university admissions departments from the rather simple gate keeping function to a diversified set of operations including marketing, recruitment, financial aid strategies, retention, and others fitting into a new set of functions called enrollment management. Higher education pricing and financial aid now occur in the context of enrollment management, an increasingly sophisticated system that addresses “the student flow” (Kroc & Hansen, 2005) from initial contact with students as early as their junior (or even 45 sophomore) years in high school through recruiting, enrollment, retention, graduation, placement, and relations with alumni (Hossler & Anderson, 2004). Enrollment management recognizes that just as students compete for admission into selective schools, institutions must compete for desirable students, where desirable usually refers to academic talent but can also apply to the goals of an appropriate balance of gender, ethnicity, socioeconomic status, or other factors such as region of the country where the student was raised. Enrollment management departments develop expensive marketing programs as elaborate as those created by Fortune 500 companies. Recruiting operations seek to maximize the pool of applications to provide the institutions the greatest possible choice of students to admit. Sophisticated computer models based on a growing body of empirical institutional research provide detailed simulations that forecast enrollment volume, statistics regarding student body composition, and expected revenue. Enrollment management is the natural product of the shift of higher education towards a competitive environment on the part of both producer and consumer. Addressing the competitive environment developing among institutions seeking to enroll desirable students, Matthew Quick wrote of the “taste for blood” and the concept of “eating another institution for lunch” to refer to one school’s stealing attractive prospects from another school (Quick, 2006). Others (Hossler, 2004; Duffy & Goldberg, 1998; Jaschik, 2006) note the concerns associated with the use of financial aid to “buy students” and the possibility that it draws both students and institutions are drawn away from the true mission of higher education. Clearly expressing anguish, in a letter to the New York Times titled, “To All the Girls I’ve Rejected,” Jennifer Britz of Kenyon University fully 46 acknowledged that she rejected female applicants with credentials superior to male applicants who were accepted (Britz, 2006). In a study examining how a public university optimizing its admissions might respond to a shock in its funding, Nagler (2008) distinguished how a school could admit affluent students known as “inferior goods” (as judged by academic merit) but still maximizing the school’s overall interest because of their ability to pay significantly higher tuition and fees. Going further than the modest enhancement of financial aid packages to “tip” student enrollment yields is the aggressive use of tuition discounts specifically designed to “buy” certain students. This is distinct from the practice of using discounts to fill empty classroom seats. The policy of filling vacant seat capacity with discounted prices has been practiced successfully by the airline industry for decades, and when an airline fills an otherwise empty chair on a plane, revenue is increased since the cost of the additional passenger is negligible. However, when a discount is used to purchase one student with desired characteristics over another student, the result can lead to institutions undercutting each other to where losses can occur. Kenneth Redd (2000) attacked the use of tuition discounting to obtain desirable students as leading to decreases in net tuition revenue to where the schools incur operating losses (Redd, 2000). While institutions may restrain themselves from such deep discounting, they do face trade-offs with respect to merit based institutional aid, in particular when it displaces funds that would have gone to students on basis of need. Ehrenberg, Zhang, and Levin (2006) studied the relationship between institution awarded National Merit Scholarships and lower-income student enrollment as quantified by those qualifying for Pell Grants. They found that on average 47 an increase of ten students receiving the merit awards was associated with a decrease of four students receiving the Pell Grants (Ehrenberg, Zhang, & Levin, 2006). Related Studies This section presents the results of some other studies relevant to the setting of tuition levels, financial aid, and the generation of revenue from sources other than tuition. Public university tuition prices and financial aid programs can vary widely from state to state. The higher education governance structures adopted by state legislatures can influence tuition prices. Some regions of the country, having greater densities of private colleges and universities, may have different political landscapes and views of what is acceptable regarding tuition and financial aid. I also explore the efforts of public universities to solicit private donations with fundraising programs, including the increasingly prevalent practice of building endowments. The most centralized governing boards, those with consolidated authority over public postsecondary institutions, have shown greater adoption of tuition “rationalization” (raising tuition to levels more closely matching the actual costs of the education) while states with coordinating boards lacking budget authority were less aggressive in raising tuition (Hearn, Griswold, and Marine, 1996). However, the significance of differences between strong and weak governing boards is declining because most states have abandoned the weak governance structure. In their study of the influence of governance structures on the management and performance of higher education institutions, Knott and Payne (2004) discovered that since 1990, eleven states had shifted to stronger governing systems, bringing the total to 37 states with strongly 48 regulated boards, with six states having moderately regulated systems and only five states with minimally regulated structures. Koshal & Koshal (2000) directly examined the relationship between state appropriation levels and tuition by analyzing a cross section of 1990 data for four-year institutions in 47 states. They found tuition to be a function of state appropriations, median family income of the state, the percentage of out-of-state students at the school, and the region of the country. The finding regarding the region of the country is consistent with the findings of increased tuition rationalization at the northeast and northern Midwest institutions, which have increased tuition substantially more than schools in the south and southwestern regions of the country (Hearn, Griswold, and Marine, 1996). In addition to setting the level of state appropriations provided to public higher education institutions, the states also set financial aid policies for the assistance provided to students. State policies shape the extent to which state resources are provided directly to institutions or directly to students, and they also influence the criteria for determining the financial aid packages students receive. These policies differ widely from state to state (Doyle, Delaney, and Naughton, 2008). As state policies have shifted from a need based focus on financial aid towards a merit based approach (Heller, 1997), what has occurred with the institutional aid at their public colleges and universities? Anticipating that perhaps the schools might use institutional aid to compensate for shifts in state aid to students, Doyle, Delaney, and Naughton (2008) specifically examined the relationship of the institutional aid packages awarded to students and the financial aid policies of the 49 states. They found that instead of compensating for shortfalls or surges due to changes in state programs, the institutions tended to reinforce or “comply” with the state’s emphasis (Doyle, Delaney, and Naughton, 2008). Such practices by the institutions only amplify the reallocation of financial aid resources from a need based orientation towards one that is merit based. One obvious choice for enhancing revenue besides increasing tuition is to pursue fundraising efforts with alumni or other community organizations. Many public universities now pursue endowments that can become as significant as those possessed by some private universities. Special projects bearing the name of a famous alumnus, a prior president, or other dignitary can solicit substantial donations. Cheslock and Gianneschi (2007) examined the ability of institutions to generate revenue with alumni fundraising and other voluntary gift programs. Their results reinforced the importance of the prestige hierarchy (as measured by selectivity), where the most selective institutions received approximately 1.8 times the gift revenue received by moderately selective schools and over 3 times the amount given to less selective institutions (Cheslock & Gianneschi, 2008). In fact, the degree of inequality along the stratification of selectivity is greater for fund raising than for state appropriations. While a nationally recognized flagship like the University Wisconsin at Madison successfully raised $1.5 billion in a fundraising campaign (University of Wisconsin Foundation, 2004), lesser known schools lack the name recognition or alumni populations to support such efforts. Examining four-year public institutions in 2004, Cheslock & Gianneschi (2008) found that a $1000 loss in state appropriations (per full time equivalent student) is associated with a $45 reduction 50 in private donations to the school. At the minimum, this casts considerable doubt on the likelihood of public universities replacing state appropriations with voluntary donations or fundraising programs. Pertinent Theoretical Frameworks Several theoretical frameworks are relevant to this study and the interpretation of its findings. The basics of economics and business administration tell us that higher education institutions must balance their budgets over time if they wish to survive. While the complex dynamics of setting tuition prices and determining financial aid packages for students produce the complicated set a relationships this study investigates, one simple fact underlies the details. As state appropriations decline as a share of the budget required to fund higher education, either other sources must fill the gap or the size of the budget must be reduced. That noted, the facts clearly show that the former (finding additional funding) is far more prevalent than the latter (reducing budgets). I have already discussed the two competing and most often discussed theories regarding the costs of higher education, the “cost disease” presented by Baumol and Blackman (1978) and Bowen’s (1980) revenue theory of cost. When examining the relationship between state appropriations and both list and net tuition prices experienced by students, the former theory would anticipate a negative relationship, with the level of state appropriations helping to alleviate the need for higher tuition prices and revenues. Bowen’s revenue theory offers a distinctly different perspective. If institutions aggressively seek to maximize revenue via all methods each optimized in its own right, then universities are already setting in-state and out-of-state pricing and packages to 51 generate the largest revenue possible. In this scenario, little or no relationship whatsoever could exist between the state appropriations made available to an institution and its pricing and financial aid policies. Looking at higher education institutions from the lens of an economist, Garvin (1980) described the university as a utility maximizing organization that seeks to increase its prestige subject to a resource constraint. Not a corporation focused solely on profits, public universities seek to improve their standing on the hierarchy of status and reputation for excellence. Financial performance remains imperative, but does so as a means to improve the school’s position compared to its competition for both students and faculty. Noting the exception of communities served by a single community college with few or no comparable alternatives in proximity, most higher education institutions compete for enrollment and prestige, but they do not all compete with each other. Rather, schools compete only with similar institutions in terms of status, geographical location, and selectivity. Nagler (2008) examined the implications of an institution that optimizes its perceived utility function on its response to funding shocks they may experience as the result of a recession or other economic calamity. What does a funding shock predict when viewed from the perspective of maximizing a function (prestige) other than profit alone? To accomplish this Nager distinguished students between those that provide prestige and those that provide revenue, since both must be considered. In such a situation, he produced a model that drew the intuitively reasonable conclusion that schools may admit “inferior goods” (referring to their academic credentials) should their 52 ability to pay out-of-state or full in-state tuition serve to sufficiently increase net tuition revenues in a response to the shock (Nager, 2008). Institutional Theory (Dimaggio & Powell, 1983) suggests that institutions of similar type (selectivity, governance) will develop highly similar practices. Their discussion of normative and mimetic institutional isomorphism provides sound explanations for the rise of relatively consistent practices and policies across institutions of similar type. As goes Harvard one can expect Stanford, Princeton, and other top privates to follow or at least seriously consider. The theory applies to public flagship institutions and community colleges as well. Such a view would predict minor variations in their response to changes in state appropriation levels per student on their pricing and aid practices. Institutional isomorphism tells us to expect the pricing and financial aid policies of similar institutions to exhibit similar behavior (Dimaggio & Powell, 1983). One could infer from this institutional theory that, for example, highly selective institutions respond in one manner while non-selective institutions react differently. Different behavior might also be found between schools located in different geographic regions, with findings for those in the northeast New England region different, perhaps substantially, from those in the South or Southwest. I also anticipate findings to differ considerably between two-year and four-year institutions. To the extent institutions are already maximizing prestige, they have tapped all resources they have identified, and a change in appropriations would have little impact on tuition. In this situation, tuition levels are influenced predominately priorities and factors (politics, student demand, etc.) that have little or nothing to do with state appropriation 53 levels. Basic economic theory would anticipate a negative coefficient for state appropriations per student when regressing listed tuition price, net tuition paid by students, or the level of unmet need. A drop in state appropriations, ceteris paribus, can be expected to associate with higher levels of list and net tuition and unmet need. A drop in state appropriations per student, ceteris paribus would also be associated with lower levels of institutional aid, since the smaller amount of state support reduces an institution’s ability to provide such grants. Conclusion As expressed by Art Hauptman (2001), the key financial components of public higher education are the state appropriations provided to institutions, the tuition levels set by those institutions, and the financial aid mechanisms in place to help aspiring students acquire the resources to pay the price. I have presented the literature discussing the declining levels of state appropriations as a portion of higher education budgets and the factors behind the decline. I have also discussed the rising tuition prices that have occurred in the last three decades and the shifting policies in financial aid during the same period. The economic conditions currently facing the United States suggest that all of these trends will continue. The questions of this study do not ask whether state appropriations will continue to decline or at what rate. Instead it seeks to clarify what data from the past suggests we can expect to happen to the in-state and out-of-state tuition levels of an institution given a particular change in its state appropriation level of support. More significantly, the study investigates how a change in state appropriations 54 translates to the tuition price and financial aid experienced by the individual students as they attempt to enroll at a public school. As demonstrated in this section, very little research has addressed this question. Furthermore, existing theories provide conflicting predictions, so we can not expect a clear answer on theoretical grounds. It is clearly an empirical question, and the next chapter will present the empirical strategy that will be utilized to answer it. 55 CHAPTER THREE DATA AND METHODOLOGY Introduction This chapter begins with the research questions that address the fundamental problem of the study, the relationship between state appropriations and higher education pricing and financial aid policies. After presenting the research questions I discuss the methodology of the study in detail, which includes the data sources, key variables, and the equations and econometric models used to conduct the analysis. After a thorough treatment of the methodology, I provide additional details on the data sources and the variables. I conclude the chapter by explaining the specific sample populations selected from the data sets and why those samples were chosen. Research Questions To clarify the relationship between state appropriations, tuition levels, and financial aid practices at public higher education institutions, I explore the following fundamental questions: 1. Given a cross section of institutional data from four-year and two-year public colleges and universities, how does the list price of tuition vary as state appropriations per student in public funding changes? a. Does the relationship between tuition and state appropriations change over the five years examined for the 1989/90 to the 2003/04 period? b. For four-year institutions, does this relationship vary between more selective, moderately selective, and less selective schools? 56 c. What is the difference in the relationships between state appropriations and in-state tuition as opposed to out-of-state tuition? d. What happens to these relationships when the five year data sets are constructed as panel data to control for institutional fixed effects? 2. When examining student-level data from a cross section of college students throughout the country, how do the list tuition, net tuition, institutional grant, and their unmet need vary as the state appropriation level per student changes? a. Does the relationship between tuition and state appropriations change over the five years examined for the 1989/90 to the 2003/04 period? b. For four year institutions, does this relationship vary between more selective, moderately selective, and less selective schools? c. How is the analysis different for students with in-state status from that of students with out-of-state status? d. What happens to the relationship between state appropriations and the items listed in question two when the student characteristics (ethnicity, gender, academic merit, etc. to be discussed in detail later) are taken into account, i.e. controlled for? 57 Methodology Overview This study combines the analysis of institutional level data for the public higher education institutions in the United States and student level data for the students attending those same institutions. For each of five separate academic years spanning the period from 1989/90 to 2003/04, I examine data for higher education institutions across the country from the Integrated Postsecondary Educational Data System (from now on referred to as IPEDS) used and maintained by the United States Department of Education. All higher education institutions in the United States are required to have their admissions or enrollment management departments collect and maintain data regarding numerous student information including enrollment levels, pricing, financial aid, demographics, and many other items. Institutions must report statistics into the IPEDS system, and detailed reports and data sets can be downloaded from the IPEDS Peer Analysis tool available on the Internet. Independently, for the same years I examine data at the student level from the National Postsecondary Student Aid Study (from now on referred to as NPSAS) also development and maintained by the United States Department of Education. The NPSAS data are collected with carefully designed surveys of a representative sample of students attending college throughout the United States. The five academic years selected for this study (1989/90, 1992/93, 1995/96, 1999/2000, and 2003/04) are those in which the NPSAS studies were conducted. I will present more about both the IPEDS data and the NPSAS data in the section on Data Sources. 58 The institutional level and student level analyses share data. For each, the same data and values are used for institutional enrollment and total state appropriations. These two numbers play the important role in the study by determining the level of state appropriations an institution receives on a per full time equivalent student basis. At the institutional level, the tuition is the value reported by the school to the IPEDS system. At the student level, the tuition is that reported for the student in the NPSAS data. The student level analysis allows for many additional variables associated with the student including ethnicity, gender, cumulative grade point average, and expected family contribution. The student level regressions control for these factors in determining the relationship between state appropriations and tuition and fees. Throughout the study, I separate the groups being analyzed into four distinct attendance scenarios consisting of 1) in-state attendance at a four-year school, 2) out-ofstate attendance at a four-year school, 3) in-state attendance at a two-year school, and 4) out-of-state attendance at a two-year school. For each of the four attendance situations, I examine the descriptive statistics of selected variables over the five years to identify trends that may be occurring. After examining trends in the descriptive statistics, I estimate ordinary least squares regressions for each individual year. This is done for the same five years for each group for both the individual and student level data sets. To prevent inflation from impacting the relationships under investigation, I convert all financial variable values to 2004 dollars using the consumer price index published by the US Department of Labor Bureau of Labor Statistics. 59 Institution Level Analysis For both four year and two year institutions, I separately run ordinary least squares linear regressions for the in-state and out-of-state listed prices for tuition and fees as reported by the institution to the IPEDS data system. I control for three factors known to influence tuition pricing at public institutions, their selectivity, the institution’s governance system, and which region of the country in which the institution resides. The econometric model consists of the following equations: In-State Tuition = β0 + β1(SA) + β2(G1) + β3(G2) + β4(S1) + β5(S2) + β6(S3) + β7(S4) + β8(NE) + β9(ME) + β10(GL) + β11(SE) + β12(SW) + β13(R) + Err Out-of-State Tuition = β0 + β1(SA) + β2(G1) + β3(G2) + β4(S1) + β5(S2) + β6(S3) + β7(S4) + β8(NE) + β9(ME) + β10(GL) + β11(SE) + + β12(SW) + β13(R) + Err SA = Total unrestricted state appropriations per FTE student G1 = Dummy variable for strong governance G2 = Dummy variable for moderate governance S1 = Dummy variable for most selective S2 = Dummy variable for more selective S3 = Dummy variable for selective S4 = Dummy variable for less selective NE = Dummy variable for New England Region ME = Dummy variable for Mid-East Region GL = Dummy variable for Great Lakes Region PL = Dummy variable for Plains Region SE = Dummy variable for Southeast Region SW = Dummy variable for Southwest Region R = Dummy variable for Rocky Mountain Region Err = Error Term The selectivity of a school serves as a powerful proxy for its status and prestige and is clearly a factor in the institution’s tuition levels and its financial aid practices 60 (Ehrenberg, 2002). The model relies on the selectivity ranking published by the 2005 Barron’s selectivity classifications. Barron’s produces a selectivity measure for institutions based upon its admission rate and the average test scores and high school rank of the students that enroll. It then groups the schools into nine categories ranging from “Most Selective” through “Non-selective.” The Barron’s groupings occur in much smaller sizes as they become more selective. To obtain groupings more suitable for this study (less disparity in sample sizes) I create five selectivity categories that combine some of the more selective Barron’s categories (this is shown in the variables section of this chapter) to produce the following groups 1) most selective, 2) more selective, 3) selective, 4) less selective, and 5) non-selective. For all descriptive statistics and regressions, an institution’s selectivity is addressed with five dummy variables, one for each selectivity category. As the regressions require the omission of a category, I omit the “non-selective” dummy variable from the regressions. While the statistics retain the variables above, when I split the institution sample into groups by selectivity, I further reduce the grouping to just three categories to obtain even greater equality in the sample sizes and simplify the analysis. The three groups used to run separate analyses are more selective (the most and more selective categories combined), middle selective (the selective category unchanged) and less selective (the less and non-selective categories combined). This applies simply to the groupings of institutions on which the analysis is performed. Selectivity is not a factor for two-year institutions, so it is not part of the model for two-year schools. 61 Studies have shown (Hearn & Griswold, 1994; Griswold & Marine, 1996; Knott & Payne, 2004) that the governance structure of the state influences the degree to they pursue policy innovations, including tuition rationalization policies. For this reason, I control for governance structure as well using dummy variables. The governance structures for the study are utilize those published by Knott & Payne (2004) and consist of 1) strongly regulated, 2) moderately regulated, and 3 ) weakly regulated. In their 2004 study, Knott and Payne started with the classification scheme of higher education governance systems from the State Postsecondary Education Structures Handbook, (1991, 1994, 1997) that distinguished board structures as consolidated board, coordinating boards, or planning agencies. Recognizing that coordinating boards could exercise different degrees of regulatory authority and seeking a classification to better measure the degree of regulation, they created three alternate classifications: highly regulated (governing board or coordinating board with strong regulation powers), moderately regulated (coordinating boards with some regulation), and weakly regulated (coordinating boards with weak regulation power, advisory boards, or planning agency). I use these classifications, and the states fitting each are listed in the variables section of these chapter. I create dummy variables for each group, and for the regressions, the omitted variable is the one indicating “weakly regulated.” In addition to selectivity and governance structure, the geographic region of a state (and embedded within it, the dominant political sentiments of the state legislatures for that state) has been shown to influence the pricing and financial aid policies of its public higher education institutions (Hearn, Griswold, & Marine, 1996). For the same 62 reasons described above, I control for the region of the country using dummy variables for each region and setting the value to one if the school is in that region. The regions match those used in the IPEDS data and consist of: 1) New England, 2) Mid-East, 3) Great Lakes, 4 ) Southeast, 5) Plains, 6) Rocky Mountains, 7) Southwest, and 8 ) Far West. The specific states in each region are listed in the variables section of this chapter. For the regressions, the omitted dummy variable is the one for “Far West.” The majority of institutions are present for all of the years, providing a panel data set on which fixed effects regression can be performed. The fixed effects regression will refine the analysis by controlling for institutional level factors that remain unchanged across the five years that are studied. Also, the fixed effects model introduces dummy variables to indicate the specific year to which the data belong. The coefficients of these dummy variables will provide robust results on the nature of tuition increases over time. Student Level Analysis The student level analysis follows the same methodology as the institutional level analysis but includes considerably more variables that are available for the students in the data, namely the characteristics of ethnicity, gender, socioeconomic status (as represented by the expected family contribution) and academic preparation or merit (as represented by the cumulative grade point average of the student at the institution). The student level study uses the same data and values for the institution’s state appropriations per student, governance, and region of the country. For all other data, the student level analysis utilizes the NPSAS variables attached to each student, including the tuition price. The student level analysis allows the study to examine the student specific information 63 regarding the net tuition paid, the institutional grant the student received, and the student’s level of unmet need. It also allows the model to control for student characteristics including ethnicity and gender, and importantly, the student’s ability to pay. As discussed in chapter two, the ability to pay is calculated from the information collected by the standardized Free Application for Federal Student Aid (FAFSA) and known as the Expected Family Contribution (EFC). The student’s ability to pay is clearly an important variable in the analysis of the financial aid awarded. The econometric model consists of the following equations: LIST TUITION = β0 + β1(SA) + β2(G1) + β3(G2) + β4(S1) + β5(S2) + β6(S3) + β7(S4) + β8(NE) + β10(ME) + β11(GL) + β12(P) + β13(SE) + β14(SW) + β15(R) + β16(W) + β17(AA) + β18(H) + β19(A) + β20(F) + β21(GPA) + β22(EFC) + Err NET TUITION = β0 + β1(SA) + β2(G1) + β3(G2) + β4(S1) + β5(S2) + β6(S3) + β7(S4) + β8(NE) + β10(ME) + β11(GL) + β12(P) + β13(SE) + β14(SW) + β15(R) + β16(W) + β17(AA) + β18(H) + β19(A) + β20(F) + β21(GPA) + β22(EFC) + Err INST GRANT = β0 + β1(SA) + β2(G1) + β3(G2) + β4(S1) + β5(S2) + β6(S3) + β7(S4) + β8(NE) + β10(ME) + β11(GL) + β12(P) + β13(SE) + β14(SW) + β15(R) + β16(W) + β17(AA) + β18(H) + β19(A) + β20(F) + β21(GPA) + β22(EFC) + Err UNMET NEED = β0 + β1(SA) + β2(G1) + β3(G2) + β4(S1) + β5(S2) + β6(S3) + β7(S4) + β8(NE) + β10(ME) + β11(GL) + β12(P) + β13(SE) + β14(SW) + β15(R) + β16(W) + β17(AA) + β18(H) + β19(A) + β20(F) + β21(GPA) + β22(EFC) + Err The Dependent Variables: List Tuition – the reported list tuition and fees prior to the application of financial aid 64 Net Tuition – the reported net tuition paid by the student after receiving financial aid (NOT including loans). Institutional Grant – the reported institutional grant provided to the student (does not have to be repaid by the student). Unmet Need – the calculated difference between the student’s ability to pay (as Measured by the expected family contribution (EFC) and the total charges incurred by the student after receiving financial aid that is NOT a loan. The Independent Variables: SA – the calculated state appropriations per full time equivalent student received by the institution. G1 – G2 – S1 – S2 – S3 – S4 – NE – ME – GL – P– SE – SW– R– AA – H– A– F– Strongly regulated governance structure Moderately regulated governance structure Most selective institution category More selective institution category Selective institution category Less selective institution category New England region Mid East region Great Lakes region Plains region Southeast region Southwest region Rocky Mountain region African American Hispanic Asian Female GPA – The student’s cumulative grade point average at the NPSAS institution EFC – The student’s reported expected family contribution Err – Error Term The purpose of the econometric model is to quantify the relationship between each of the dependent variables and the state appropriations per student received by the institution. The other variables are included as control variables to account for other factors that may be influencing the values of the dependent variables. 65 The NPSAS survey incorporates weighting and sample strata to insure that minority populations are sampled in sufficient quantity to obtain meaningful results. This weighting can influence both the descriptive statistics and the results of the regression, so I run the analysis both with and without the incorporation of the sample design. I run the analysis first without any weighting or sample design, then with the weighting variable alone, and finally with the full sample design taken into consideration. For all regressions, the analysis is run on those students for which the data is complete, resulting in slightly different sample sizes. Data Sources Institutional Level Analysis For the institutional level data, this study uses data from the United States Department of Education’s National Center for Education Statistics Integrated Postsecondary Education Data System (IPEDS) for the schools years 1989-1990, 19921993, 1995-1996, 1999-2000, and 2003-2004. Federal law requires all higher education institutions to report detailed data regarding their operations including student enrollment numbers for full and part-time students at each grade level. In addition to enrollment data, schools must report detailed financial information that include prices for tuition, fees, room and board, estimates for other expenses and with the detail necessary to distinguish in-district, in-state, and out-of-state pricing. Importantly, all institutions must report the amount of state appropriations they receive each year, a value fundamental to this study. IPEDS data are available to the public, and for this study the pertinent data sets were downloaded as spreadsheets from the IPEDS Peer Analysis System Web site 66 hosted by the National Center for Education Statistics. The specific data sets downloaded are listed in detail in Appendix A. After the spreadsheets were stored on the local hard drive, the statistical software STATA 10.0 was utilized to perform the analyses. The IPEDS data set does not include the selectivity of an institution. Throughout the study, for both the institutional level and student level analyses, four-year institutions were assigned a selectivity value based upon their designation in the 2005 Barron’s Profiles of American Colleges. As noted in the previous section, for the level of regulation exerted by the governing boards of the states in which the institutions reside, it was also necessary to use another source. I use the categories as published by Knott & Payne (2004) that classifies states as 1) minimally regulated, 2) moderately regulated, and 3) highly regulated. As I discussed in the literature review, the state governance system was found to be significant in several studies on tuition rationalization. Since so many states have adopted stronger governance arrangements in recent years (Knott & Payne, 2004), I anticipate a decline in variation explained by governance structure but consider its inclusion desirable for consistency with other studies and confirmation of the decline. The list of states in each category is in the variables section of this chapter. Student Level Analysis For the student level analysis, I use the United States Department of Education’s National Center for Education Statistics National Postsecondary Student Aid Study (NPSAS) for the same years as above, 1989-1990, 1992-1993, 1995-1996, 1999-2000, and 2003-2004. The NPSAS study is designed to select a nationally representative 67 sample of postsecondary education institutions. NPSAS obtains and fortifies its data by using multiple sources including national databases, computer assisted telephone methods (CATI), institutional records, and student interviews. The student level study utilizes the same measures for institution selectivity and state governance structure as that described for the institutional level. Variables Variables in both the Institution and the Student Level Analyses The level of state appropriations received by an institution plays a central and critical role in the entire study and warrants a detailed explanation. The value is produced entirely with data from the IPEDS system. I directly read the value of total state appropriations received by each institution from the IPEDS financial data set that contains this variable for each institution. To calculate the state appropriations received per full time equivalent student, it is necessary to produce a variable that expresses the total enrollment of full time equivalent students. The total enrollment at the school is calculated to full time equivalent students using the following formula: Total FTE students = (total full time undergraduate students) + 0.333 x (total part time undergraduate students) + 2 x (total full time graduate students + 0.666 (total part time graduate students) The formula has been used by Cheslock and Giannaschi (2008) and in other studies (Bowen, 1980; Brinkman, 1990) to account for the additional expenses associated with graduate students and reduced costs for part time students. The total appropriations received at the institutions are divided by the total full time equivalent enrollment to generate the variable used in the regression, state appropriations received per student. 68 State appropriations per student = (Total state appropriations) (Total calculated FTE) The selectivity dummy variables are generated from the classifications assigned to the schools in the 2005 Barron’s publication. The Barron’s publication places greater ranking sensitivity of its grouping for the more selective schools (it splits the top 7% of schools into three groups while the bottom 12% are one group). To produce groups better suited to this study, I collapse the more selective Barron’s groups as shown below. The percentage specified in parentheses indicates the portion of schools that fall into the category. Selectivity Dummy Variables: Select1 = Most competitive + Highly competitive plus + Highly competitive (7%) Select2 = Very competitive plus + Very competitive + competitive plus (20%) Select3 = Competitive (43%) Select4 = Less competitive (18%) Select5 = Non competitive (12%) For the governance classifications, three dummy variables are created. G1 – Highly Regulated Alabama, Arizona, Arkansas, Connecticut, Colorado, Indiana, Florida, Georgia, Idaho, Illinois, Iowa, Kansas, Kentucky, Louisiana, Maine, Maryland, Massachusetts, Mississippi, Missouri, Montana, Nevada, New Hampshire, North Carolina, Ohio, Oklahoma, Oregon, Rhode Island, South Carolina, South Dakota, Tennessee, Utah, Wisconsin, Wyoming G2 – Moderately Regulated Nebraska, New Jersey, New York, Texas, Virginia, Washington G3 – Minimally Regulated California, Delaware, Michigan, New Mexico, Pennsylvania. I adopt the regional classifications used by the IPEDS data. This system classifies the states as follows: 69 The New England Region – Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island and Vermont. The Mid-East Region – Delaware, District of Columbia, Maryland, New Jersey, New York, Pennsylvania The Great Lakes Region – Illinois, Indiana, Michigan, Ohio, Wisconsin The Plains Region – Iowa, Kansas, Minnesota, Montana, Nebraska, North Dakota, South Dakota The Southeast – Alabama, Arkansas, Florida, Georgia, Kentucky, Louisiana, Mississippi, Tennessee, Virginia, West Virginia The Southwest – Arizona, New Mexico, Oklahoma, Texas The Rockies – Colorado, Idaho, Montana, Utah, Wyoming The Far West – California, Alaska, Hawaii, Nevada, Oregon, Washington Institutional Level Analysis Variables For the institution level regressions, the primary dependent variable is the listed price for tuition and fees as reported in IPEDS with the primary independent variable of state appropriations per student. Four-year institutions are run separately from two-year institutions, and in-state prices are distinguished from out-of-state prices. IPEDS TUITION: The listed tuition prices are directly available from the IPEDS institutional characteristics data sets and are available for in-district, in-state, and out-ofstate status. The specific variable names accessed in each data set are listed in Appendix A, which includes all IPEDS data sets utilized with the variable names read from each. It is worth noting that IPEDS collects tuition data in a variety of ways including charges by credit hour, charges for different programs, graduate versus undergraduate charges, including tuition associated fees (or not including them), and so on. For this study the variable selected is that which represent full time full year annual tuition and fees for an undergraduate student. 70 Student Level Analysis Variables The student level analysis adopts the same institutions and institutional characteristics utilized in the institutional level analysis. The same data and calculation for state appropriations per student is used as is the criteria for including the institution (and therefore, the students attending that institution) in the study. In addition to these, the student level data allow for far more precise regressions controlling for many student characteristics and financial aid information. Four measures are run separately as the dependent variable (all as reported by NPSAS), listed tuition price, net price paid by the student, institutional grants, and the student’s unmet need. The primary independent variable is the state appropriations per student received by the school, but many other independent variables are included as controlling variables including the student’s ethnicity, gender, academic merit, and socioeconomic status. This allows for a more precise regression of the relationships. NPSAS TUITION: Like the approach for the IPEDS data, the variables used for listed tuition price is the variable corresponding to the full year full time undergraduate tuition. Reported on individual students, the variable is the same whether the student is in-state or out-of-state, a status that is reported with the jurisdiction variable of the data set. In instances where the tuition prices reported by the institution are different from the tuition prices reported for individual students at that same institution, this analysis uses the student data. Why the prices would be different is outside the scope of this work, and efforts to understand this continue. 71 NET TUITION is the list tuition reported for the student minus the institutional grant the student received. In this study, the student’s unmet need is the difference between a student’s calculated expected family contribution and the costs actually incurred after financial aid not including loans. The intent behind the variable is to measure the degree of financial burden, including the debt incurred, associated with paying for college. I anticipate this result to be more statistically significant and compelling than the analysis of tuition levels. GPA: The student’s cumulative grade point average refers to that of the student at the institution in which the student is enrolled for the NPSAS sample. This is NOT the student’s high school grade point average prior to admission. For this study, its value is used as a proxy for academic merit. INSTITUTIONAL GRANT – This variable is reported directly in the NPSAS data. STUDENT CHARACTERISTICS – The NPSAS data contain the student specific characteristics of ethnicity, gender, age, dependency status, attendance pattern, and importantly, a great deal of financial information. For this study, the composite expected family contribution (EFC) variable is used. The composite variable is the most robust in the NPSAS data set, with considerable effort on the part of NPSAS management to impute values when necessary. 72 Sample Populations Institution Level Analysis The sample is restricted to public non-profit institutions that enroll undergraduate students. The sample was created by starting with the full universe of public institutions that reported to IPEDS and then dropping institutions that did not receive any state appropriations, institutions that are labeled as less-than two-year institutions, and observations that were listed as district offices, administration headquarters, or other overhead associated entities that do not have students. Other cases were dropped due to missing data on institutional enrollment or selectivity. All private institutions including proprietary schools are excluded. The descriptive statistics for the variables used in the institutional level analysis are provided in Appendix Table 7. TABLE 2. The IPEDS Institution Sample Populations Academic Year 2003 – 2004 1999 – 2000 1995 – 1996 1992 – 1993 1989 – 1990 Four Year Institutions 432 463 460 468 453 Two Year Institutions 900 941 797 864 858 Student Level Analysis For the student level analysis with the NPSAS data, the study merges the studentlevel NPSAS data with institutional data from IPEDS and other sources. To maximize the comparability of the two analyses, the student sample is restricted to those at the same 73 institutions in the IPEDS study. The students are then restricted further to undergraduates that are dependent, full-time, full year students who attended a single institution during the year of the NPSAS sample. The NPSAS Student Sample Populations The NPSAS population consists of full time full year dependent undergraduates attending institutions included in the IPEDS institution population. Table Three shows the sample populations for in-state and out-of-state students at the institutions for each academic year. It also shows the sample sizes for the institutions disaggregated into the three selectivity subgroups. The descriptive statistics of the key variables used in the analysis are provided in Appendix Table 8. 74 Table 3 The NPSAS Sample Population Sizes Four Year Institutions (N = Number of Students) 1989 – 1990 All schools More Selective Middle Selective Less Selective In-State Students 3352 1177 1714 461 Out-of-State Students 395 179 159 57 1992 – 1993 All schools More Selective Middle Selective Less Selective In-State Students 5433 2517 2180 736 Out-of-State Students 1233 633 425 175 1995 – 1996 All schools More Selective Middle Selective Less Selective In-State Students 5533 2349 2222 961 Out-of-State Students 1161 613 435 113 1999 – 2000 All schools More Selective Middle Selective Less Selective In-State Students 5863 2577 2464 822 Out-of-State Students 1039 600 365 74 2003 – 2004 All schools More Selective Middle Selective Less Selective In-State Students 7669 2880 3679 1101 Out-of-State Students 1061 527 421 111 75 Selectivity is not a factor at community colleges, so no disaggregation of schools by selectivity is performed. The NPSAS samples in the earlier years have small representation of two-year schools. While I include the analysis of two-year out of state students for sake of completeness, the sample sizes are extremely small and the results found for this group should be interpreted with caution. Table Four provides the sample sizes for each year. Table 4 NPSAS Sample Population Sizes Two-Year Institutions (N = Number of Students) Year 1989 – 1990 1992 – 1993 1995 – 1996 1999 – 2000 2003 – 2004 In-State Students 267 398 489 693 3156 Out-of-State Students 39 37 28 102 386 76 CHAPTER FOUR RESULTS Introduction I present the results in two parts, the first part consisting of the institutional level analysis using the IPEDS data. The second part consists of the student level data obtained from the NPSAS data. For both sections, I first present the descriptive statistics and discuss the trends and other key findings. Then I show the results of the ordinary least squares regressions. Throughout the study, the analysis is conducted separately for the four different attendance scenarios: in-state and out-of-state status at four-year schools, and in-state and out-of-state status at the two-year schools. All financial numbers are expressed in 2004 dollars as calculated by the Consumer Price index published by the Bureau of Labor Statistics. After presenting the results for the institutional level and student level models, I discuss the similarities and differences found between the two and explain the factors that might be causing them. Recall from chapter three that the selectivity of the four-year institutions is based on the Barron’s classifications but collapsed to five categories. For more desirable sample sizes, the schools were further combined to produce three populations, more selective (the first two combined), middle selective (the third group left as is), and less selective (the final two combined). 77 Institution Level Analysis – Four-Year Institutions Table Five provides the summary statistics on the both the in-state and out-ofstate tuition charged at four-year institutions over the five academic years. It shows that the data examined contain a high degree of consistency from year to year. The sample populations consist of essentially the same set of institutions, varying from 432 to 468 schools. The selectivity categories are also very consistent in terms of size and data relationships. The data strongly demonstrate the positive relationship between school selectivity and tuition price for both in-state and out-of-state pricing. Average in-state tuition at the more selective schools exceeded that for the middle selective schools by approximately 20%, and difference in tuition from the middle selective and the less selective is about 12%. Selectivity plays a stronger but also apparently stable role with out-of-state tuition. For out-of-state tuition, the more selective institutions on average charge tuition at about 25% more than those in the middle selective category, which charge about 10% more than those in the less selective group. For all four-year institutions, in-state tuition prices far outpaced the rate of inflation, climbing from an average of $2456 to $4443 or 81% over the fourteen year period. Out-of-state tuition rose from an average of $6462 to $11,211, an increase of 73.5%. The data also confirm the literature regarding the stagnation of state appropriations received per student over the period of the study. The data for the total sample show average state appropriations varying from a high of $6727 in 1989/90 to a 78 TABLE 5. Institutional Level Analysis – Four Year Institution Descriptive Statistics Integrated Postsecondary Education Data System (IPEDS) Sample Size Mean / Std. Deviation All Institutions More Selective Middle Selective Less Selective 1989 – 1990 In-State Tuition Out-of State Tuition State Approp. Per Student (N=453) 2456 / 883 6462 / 2110 6727 / 2859 (N = 120) 2822/ 1020 7861/ 2385 8375 / 3111 (N = 196) 2427 / 840 6180 / 1767 6370 / 2641 (N = 134) 2163 / 679 5655 / 1682 5828 / 2260 1992 – 1993 In-State Tuition Out-of State Tuition State Approp. Per Student (N=468) 2935 / 1059 7689 / 2565 5798 / 2409 (N = 128) 3441/ 1229 9344 / 2873 7156 / 2201 (N = 202) 2881/ 972 7396 / 2178 5519 / 2200 (N = 134) 2517 / 787 6616 / 1964 5020 / 1786 1995 – 1996 In-State Tuition Out-of State Tuition State Approp. Per Student (N=460) 3278 / 1147 8759 / 2723 5923 / 2237 (N = 124) 3929 / 1350 10549 / 2798 7047 / 2557 (N = 201) 3172 / 1020 8382 / 2331 5683 / 2028 (N = 134) 2819 / check 7689 / 2182 5275 / 1799 1999 – 2000 In-State Tuition Out-of State Tuition State Approp. Per Student (N=463) 3483 / 1154 9464 / 2685 6581 / 2448 (N = 127) 4159/ 1335 11446 / 2943 7804 / 2926 (N = 201) 3368 / 1011 9038 / 2170 6234 / 2089 (N = 134) 3019 / 852 8233 / 2028 5939 / 2029 2003 – 2004 In-State Tuition Out-of State Tuition State Approp. Per Student (N=432) 4443 / 1495 11211 / 3407 5595 / 2323 (N = 114) 5334 / 1869 13791 / 3820 6671 / 2708 (N = 187) 4302 / 1217 10793 / 2586 5312 / 2172 (N = 127) 3848 / 1100 9603 / 2731 5120 / 1811 79 low of $5598 in 2003/04 with levels rising and falling. For all four year institutions, a trend line over period of the study declines with a slope of $40 (negative) per year in state appropriations received per student. The level of state appropriations received per student is also stratified by selectivity, with more selective institutions on average receiving about 25% more than the middle selective schools, but the difference in state appropriations received by middle selective schools and the less selective is rather small, only 4% in 2003/04. Summarizing, the key conclusions to be drawn regarding tuition from the descriptive statistics are that both in-state and out-of-state tuition are rising far faster than the rate of inflation and that the stratification of pricing along the selectivity of the institution is increasing in actual dollars, growing in a way that approximately maintains the differences as ratios or expressed as percentages. State appropriations have stagnated but remain stratified with the more selective schools receiving more funding per student, although the difference between the middle selective and the less selective schools is diminishing, only 4% between them in 2003/04. Table Six shows the results of the ordinary least squares regressions performed on the institutional data for in-state tuition pricing. The data are consistent with the findings in the literature (Koshal & Koshal, 2000; Hossler et al., 1997) regarding the suppressing influence of state appropriations on in-state tuition levels. The findings do indicate a negative relationship between state appropriations and in-state tuition levels, as negative coefficients were found in almost every regression. For the total sample the coefficient ranged from – 0.043 (a $100 drop in state appropriations per student is 80 Table 6: Institutional Level Analysis – Four Year Institution Regression Integrated Postsecondary Education Data System (IPEDS) In-State List Tuition dependent on State Appropriations 1989 – 1990 All Institutions More Selective Middle Selective Less Selective B - 0.043 - 0.046 - 0.018 - 0.080 Std. Dev. 0.013 0.027 0.020 0.018 t value - 3.13 - 1.71 - 0.90 - 4.30 P>|t| 0.001 0.091 0.371 0.000 1992 – 1993 All Institutions More Selective Middle Selective Less Selective - 0.031 - 0.031 - 0.007 - 0.082 0.016 0.031 0.024 0.025 - 1.96 - 0.99 - 0.28 - 3.29 0.051 0.324 0.783 0.001 1995 – 1996 All Institutions More Selective Middle Selective Less Selective - 0.026 - 0.023 - 0.013 - 0.076 0.017 0.035 0.024 0.023 - 1.59 - 0.67 - 0.56 - 3.31 0.112 0.505 0.579 0.001 1999 – 2000 All Institutions More Selective Middle Selective Less Selective - 0.015 - 0.008 0.002 - 0.079 0.016 0.031 0.023 0.024 - 0.97 - 0.26 0.10 - 3.33 0.331 0.792 0.923 0.001 2003 – 2004 All Institutions More Selective Middle Selective Less Selective - 0.014 - 0.032 - 0.001 - 0.042 0.025 0.061 0.029 0.043 - 0.058 - 0.52 - 0.04 - 0.98 0.565 0.602 0.967 0.327 81 associated with a $4.30 increase in tuition) in 1989/90 to – 0.014 (a tuition increase of only $1.40). Interestingly, the magnitude of the coefficients for each year falls rather steadily as follows: - 0.043, - 0.031, - 0.026, - 0.015, - 0.014. What also occurs is a steady decline in the significance of the results, starting at a highly significant p value of 0.001 for 1989/90, then 0.051, 0.112, 0.331, and 0.565 respectively for the following years. One interpretation of these two trends is that the relationship between state appropriations and in-state tuition across institutions inside of a given year used to exist as discussed in Koshan & Koshan (2000) but has since essentially disappeared. Another explanation is that while state appropriations remain a critical funding stream, as they decline in significance as a part of institutional budgets, appropriations are playing less of a role (when comparing institutions) in determining tuition levels. Examining the institutions within the different levels of selectivity shows that the less selective schools for every year exhibit a stronger relationship between state appropriations and tuition prices. With the exception of 2003/04, at the less selective schools a $100 increase in tuition is associated with a tuition increase of approximately $8.00, double the value of the other schools and with very significant p values of 0.001 or 0.000. While the less selective schools show the strongest response, the middle selective schools show the least. The data show no virtually no relationship at all between state appropriations and tuition at the schools in the middle selective category for any of the years. Table Seven shows the regression coefficients for out-of-state tuition at four-year institutions. The findings show a positive relationship between state appropriations and 82 tuition levels. I believe this is the result from a positive relationship between prestige, state appropriations, and out-of-state tuition. The out-of-state tuition regressions over the five years showed a high level consistency in the coefficient for the state appropriations variable. For all five years, it remained within the range of 0.047 (2003/04) to 0.071 (1992/93), steadily declining from 1992/93 to its value in 2003/04. Like the results for in-state tuition, one clear trend is the decline in the statistical significance of the coefficients. Over the five years the significance of the relationship steadily decreased, with the associated p values climbing as follows: 0.073, 0.092, 0.172, 0.250, and 0.437. This suggests that for both in-state and out-of-state tuition, the association with state appropriations (across institutions in a single year) is falling during the period of the study. When looking at the results with the schools disaggregated by selectivity, no consistent picture emerges that remains constant for the five years. In the first two years, the strength of relationship between state appropriations and tuition correlated with selectivity with the most selective institutions have the highest coefficient, followed by the middle selective and then the less selective schools. However, this situation reverses in 1995/96 and by 2003/04, the least selective institutions had the largest coefficient, one that suggests a $19.10 increase in tuition is associated with a $100 increase in state appropriations. 83 TABLE 7: Institutional Level Analysis – Four Year Institution Regression Integrated Postsecondary Education Data System (IPEDS) Out-of-State List Tuition dependence on State Appropriations B Std. Dev. t value P>|t| 1989 – 1990 All Institutions More Selective Middle Selective Less Selective 0.055 0.071 0.062 0.044 0.031 0.059 0.045 0.058 1.80 1.19 1.37 0.76 0.073 0.235 0.172 0.446 1992 – 1993 All Institutions More Selective Middle Selective Less Selective 0.071 0.081 0.072 0.044 0.042 0.074 0.062 0.081 1.69 1.09 1.17 0.55 0.092 0.277 0.245 0.586 1995 – 1996 All Institutions More Selective Middle Selective Less Selective 0.062 0.038 0.043 0.119 0.045 0.082 0.063 0.085 1.37 0.47 0.68 1.40 0.172 0.641 0.498 0.165 1999 – 2000 All Institutions More Selective Middle Selective Less Selective 0.053 0.027 0.085 0.013 0.046 0.084 0.069 0.090 1.15 0.32 1.23 0.14 0.250 0.753 0.221 0.886 2003 – 2004 All Institutions More Selective Middle Selective Less Selective 0.047 0.049 0.009 0.191 0.060 0.117 0.084 0.122 0.78 0.42 0.10 1.56 0.437 0.674 0.917 0.122 84 While the relationships between tuition and state appropriations were weak, the regressions did explain a considerable amount of the variation in tuition levels. The full regression analyses showing the coefficients for all of the independent variables are provided in Appendix Table 1 (in-state tuition) and Appendix Table 2 (out-of-state tuition). Summarizing, the adjusted R squared values ranged from 0.4397 to 0.6018, suggesting that the variables explained approximately half of the variation in tuition prices. The results confirm the literature that the New England and Mid East regions have significantly higher tuition prices followed by the Great Lakes and Plains regions, with the Southeast and Southwest regions having the lowest tuition. Not surprisingly, the dummy variable associated with a school’s being in the most selective category had the largest of the selectivity variable coefficients, ranging between $1200 and $1400 until 2003/04, where this coefficient jumped to $2143. The coefficients for the selectivity variables decrease as one drops from a higher to a lower category. The variable for strongly regulated governance systems added between $255 (1999/00) and $677 (1995/96) to in-state tuition, a finding also consistent with the literature that ties stronger regulation to increased tuition rationalization. . Institution Level Analysis – Two-Year Institutions Table Eight contains the descriptive statistics for the two-year institutions which also distinguish in-district tuition from in-state and out-of-state tuition, and for all years where in-district tuition was available (the IPEDS data for 1999/2000 did not contain an in-district tuition variable) it was lower than in-state tuition by approximately $250 to 85 $300 per year (a substantial 20 - 30%). As with the four-year schools, tuition is increasing at rates far exceeding inflation, rising 57.7%, 50.0%, and 37.6% for in-district, in-state, and out-of-state respectively over the fourteen year period. State appropriations have also stagnated for the two-year schools, rising and falling but remaining in the vicinity of $4450 per student. TABLE 8: Institution Level Analysis – Two Year Institution Descriptive Statistics: Tuition and State Appropriations Year and Sample Size Mean / Standard Error 1989 – 1990 (N = 858) 1992 – 1993 (N = 864) 1995 – 1996 (N = 797) 1999 – 2000 (N = 941) 2003 – 2004 (N = 900) In-District Tuition In_State Tuition Out-of-State Tuition State Approp / FTE 1221 / 721 1438 / 784 1591 / 803 NA 1925 / 884 1546 / 1096 1752 / 1088 1811 / 1014 1609 / 1045 2319 / 1293 3843 / 882 4431 / 2080 4597 / 2275 4247 / 2313 5287 / 2484 4456 / 2335 3862 / 4921 4437 / 2367 5297 / 9494 4481 / 10691 While the sample sizes of the two-year schools were larger than the four-year institution samples, ranging from 797 to 941, the two-year schools show a far higher degree of variation in their tuition levels. While the samples of four-year schools had standard deviations of about a third of the mean values, the samples of the two-year institutions have standard deviations closer to a half of the mean values. The degree of variation among two-year schools is even greater for the level of state appropriations they receive, and the data suggest the variation is increasing. For the year 1999/2000, state appropriations per full time equivalent student averaged $5297 with a standard deviation of $9494. For 2003/2004, the mean funding per student fell to $4481 while the standard 86 deviation rose to $10,691. This degree of variation suggests that the universe of two-year colleges consists of substantially different financing and pricing environments across the country. These differences could also be due to more variation in calculating the full time student equivalent enrollment at the two-year institutions. Table Nine shows the results of the regression for the two-year institutions including in-district, in-state, and out-of-state tuition. For all dependent variables, the coefficients of the independent state appropriations per student variable are very small and most have minimal statistical significance. Given the relatively small tuition levels at two-year institutions and their reliance on local appropriations as well as state appropriations, these results are not surprising. 87 TABLE 9: Institution Level Analysis – Two Year Institution Regressions State Appropriation Coefficient on Tuition Prices Tuition Price on Appropriations B Std. Dev. t value P>|t| 1989 – 1990 (N = 858) In-District In-State Out-of-State 0.014 - 0.053 0.032 0.008 0.011 0.026 1.76 - 4.96 1.26 0.078 0.000 0.210 1992 – 1993 (N = 864) In-District In-State Out-of-State 0.005 - 0.008 0.011 0.004 0.005 0.012 1.47 - 1.75 0.89 0.142 0.080 0.371 1995 – 1996 (N = 797) In-District In-State Out-of-State 0.013 - 0.034 - 0.001 0.007 0.010 0.030 1.73 - 3.40 -0.04 0.084 0.001 0.969 1999 – 2000 (N = 941) In-District In-State Out-of-State NA 0.001 - 0.013 NA 0.002 0.007 NA 0.36 - 1.74 NA 0.717 0.083 2003 – 2004 (N = 900) In-District In-State Out-of-State 0.004 0.002 0.001 0.002 0.003 0.007 2.18 0.56 0.13 0.030 0.579 0.897 The complete regression results for two-year institutions are shown in Appendix Table 3. As with the four-year institution regressions, even though the relationship between tuition and state appropriations was weak, the regressions explained a good deal of the variation in price (slightly more than the four-year regressions) with adjusted R squared values over 0.6 for every year but 2003/04, where the value was 0.5254. Although not in the same proportions, the results are approximately consistent with the four-year schools regarding the region of the country with the exception that the Mid East and Plains regions both surpass New England in tuition prices. While not a central focus 88 of the study, the results for governance are quite interesting as the sign for the moderately and strongly regulated governance is negative for the two-year institutions. Summarizing, for purposes of this study, the key point from the institutional level regressions above is that within a given year, state appropriations play a small role, pennies on the dollar, in explaining the variation of in-state tuition across institutions. The in-state analysis found a negative relationship that steadily decreases in magnitude from – 0.043 to – 0.014. For out-of-state tuition, the relationship is positive, suggesting that institutions with higher state appropriations may have more prestige and charge higher prices to nonresident students. For both in-state and out-of-state tuition, the results show declining statistical significance. The above cross-section regressions focused on examination variation across institutions at a point in time. In other words, the results were essentially identified by comparing tuition levels for institutions in high state appropriations states with tuition levels for institutions in low state appropriations states. But it would be more interesting to examine how the tuition of individual institutions varied over time as their levels of state appropriations were altered. For this reason, the study conducts a fixed effects institutional level regression that simultaneously includes data for all five years. As shown in Table 10, starting with the four-year institutions, the fixed effects results match the negative relationship between state appropriations and in-state tuition, but show a stronger link, suggesting that a $100 decrease in state-appropriations is associated with a $7.10 increase in in-state tuition, seven cents on the dollar. When examining the results 89 Table 10 Institutional Level Analysis Fixed Effects Regression Four Year Institutions All Institutions In-state Tuition B State Appropriations Year 9293 Year 9596 Year 9900 Year 0304 - 0.071 580 1087 1619 2860 Out-of-state Tuition B State Appropriations Year 9293 Year 9596 Year 9900 Year 0304 - 0.069 1446 2855 4257 7041 Std. Dev. t value P>|t| 0.014 32.3 32.8 37.9 77.8 - 4.95 17.97 33.12 42.71 77.79 0.000 0.000 0.000 0.000 0.000 Std. Dev. t value P>|t| 0.033 74.8 76.0 87.8 85.1 - 2.09 19.33 37.57 48.48 82.71 0.036 0.000 0.000 0.000 0.000 Std. Dev. t value P>|t| 0.030 82.7 81.8 91.8 91.0 - 2.28 9.59 17.33 22.13 39.58 0.023 0.000 0.000 0.000 0.000 Std. Dev. t value P>|t| 0.061 164 166 186 185 - 1.51 11.00 20.65 27.86 47.64 0.133 0.000 0.000 0.000 0.000 More Selective Institutions In-state Tuition B State Appropriations Year 9293 Year 9596 Year 9900 Year 0304 - 0.068 774 1418 1619 3602 Out-of-state Tuition B State Appropriations Year 9293 Year 9596 Year 9900 Year 0304 - 0.091 1802 3431 5190 8803 90 Table 10 Institutional Level Analysis Fixed Effects Regression on Five Year Panel Data (continued) Four Year Institutions Middle Selective Institutions In-state Tuition B State Appropriations Year 9293 Year 9596 Year 9900 Year 0304 - 0.096 550 1026 1554 2739 Out-of-state Tuition B State Appropriations Year 9293 Year 9596 Year 9900 Year 0304 - 0.151 1411 2779 4201 6853 Std. Dev. t value P>|t| 0.019 39.5 40.1 46.7 44.92 - 5.08 13.90 25.59 33.25 60.98 0.000 0.000 0.000 0.000 0.000 Std. Dev. t value P>|t| 0.045 95.0 96.4 112 108 - 3.34 14.85 28.84 37.40 63,50 0.001 0.000 0.000 0.000 0.000 Std. Dev. t value P>|t| 0.021 43.4 44.5 53.8 51.1 - 2.12 10.64 19.83 25.68 46.88 0.035 0.000 0.000 0.000 0.000 Std. Dev. t value P>|t| 0.059 120 123 147 142 1.12 10.02 19.88 23.94 40.80 0.262 0.000 0.000 0.000 0.000 Less Selective Institutions In-state Tuition B State Appropriations Year 9293 Year 9596 Year 9900 Year 0304 - 0.045 462 881 1354 2395 Out-of-state Tuition B State Appropriations Year 9293 Year 9596 Year 9900 Year 0304 0.067 1207 2452 3511 5785 91 Table 10 Institutional Level Analysis Fixed Effects Regression on Five Year Panel Data (continued) Two Year Institutions In-state Tuition B State Appropriations Year 9293 Year 9596 Year 9900 Year 0304 - 0.001 278 511 869 1116 Out-of-state Tuition B State Appropriations Year 9293 Year 9596 Year 9900 Year 0304 0.003 833 1326 1873 2816 Std. Dev. t value P>|t| 0.003 20.8 21.5 21.3 21.2 - 0.22 13.37 23.75 40.83 52.56 0.822 0.000 0.000 0.000 0.000 Std. Dev. t value P>|t| 0.007 74.8 76.0 87.8 85.1 0.47 19.33 37.57 48.48 82.71 0.637 0.000 0.000 0.000 0.000 92 for institutions at the different levels of selectivity, the middle selectivity group had the largest response, a $9.60 increase compared to $6.80 for the more selective and $4.50 for the less selective. Curiously, the fixed effects results for out-of-state tuition paint the opposite picture of the single year regressions, showing that a $100 decrease in state appropriations is associated with a $6.90 increase in tuition (the opposite direction), almost identical to the seven cents on a dollar result for in-state tuition. In the fixed effects analysis, the increasing coefficients for the year indicator variables are very strong and reinforce the rapid rise in tuition prices over the period of the study. For the two-year schools, the fixed effect model matched the results of the prior regressions, also finding no significant relationship between state appropriations and either the in-state or the out-of-state tuition prices. The coefficients for the state appropriations variables are extremely small and have negligible statistical significance, with p values exceeding 0.6 for both in-state and out-of-state regressions. As with the four-year schools, the increasing coefficients for the year indicator variables indicate the rise in tuition prices over time. 93 Student Level Analysis - Four-Year Institutions One would anticipate slight differences to exist between the institution level data and the student level data since they come from different data sets and are reporting summary statistics on different populations (institutions vs. students). However, the two should have a degree of correspondence, and they do in the case of the descriptive statistics. With the exception of 1989/90, the institutional level average in-state tuition at the four-year institutions tracks with the student level average in-state tuition at four-year schools, with the student data consistently averaging in-state tuition prices approximately $300 – $600 higher than that reported by the institutions. The out-of-state comparison differs in that the students reported substantially lower tuition prices for 1989/90 ($4339 compared to $6462 for the institutions), about the same in 1992/93 ($7495 compared to $7689 for the institutions), and then reporting higher prices by about $1200 for the last three years. As shown in Table 11, the student level data also show the state appropriations per student not only stagnating but declining. Like tuition, state funding at an institution for in-state students and out-of-state students are relatively close within a given year. The in-state tuition, out-of-state tuition, and state appropriations per student means are consistently higher for the student level data. This result reflects the greater representation of students from high-tuition and high-appropriations schools in the NPSAS samples. 94 TABLE 11: Student Level Analysis – Descriptive Statistics: Four Year Institutions Sample Size Mean / Std. Deviation All Institutions More Selective Middle Selective Less Selective 1989 – 1990 In-State Students Tuition Institutional Grant Unmet Need State App per FTE (N=3570) 1675 / 1400 149 / 703 1720 / 4196 7227 / 3052 (N = 1270) 1869 / 1889 187 / 757 1855 / 4582 8658 / 3049 (N = 1795) 1563 / 978 123 / 683 1614 / 3920 6462 / 2810 (N = 505) 1541 / 991 136 / 607 1734 / 4051 6023 / 2293 (N = 431) 4339 / 2678 222 / 987 2378 / 6276 7108 / 2540 (N = 201) 5139 / 3038 150 / 981 2747 / 6563 7950 / 2606 (N = 162) 3608 / 1994 267 / 998 2207 / 6180 6582 / 1967 (N = 68) 3400 / 1952 363 / 969 1510 / 5410 5480 / 2512 (N = 5916) 3204 / 1529 346 / 1152 5449 / 2706 6568 / 2535 (N = 2831) 3496 / 1799 379 / 1255 5689 / 2869 7640 / 2708 (N = 2284) 3006 / 1193 340 /1086 5241 / 2365 5685 / 1932 (N = 801) 2719 / 1029 247 / 923 5372 / 3080 5222 / 1643 (N = 1374) 7495 / 4370 1069 / 2895 6426/ 5065 6181 / 2374 (N = 707) 9008/ 4729 1116 / 3032 7892 / 5434 6731 / 2577 (N = 467) 6154 / 3325 1036 / 2765 5117 / 4194 5759 / 2402 (N = 200) 4694 / 2237 958 / 2643 3736 / 3234 4963 / 1538 Out-of State Students Tuition Institutional Grant Unmet Need State App per FTE 1992 – 1993 In-State Students Tuition Institutional Grant Unmet Need State App per FTE Out-of State Students Tuition Institutional Grant Unmet Need State App per FTE 95 TABLE 11: Student Level Descriptive Statistics (continued) Sample Size Mean / Std. Deviation All Institutions More Selective Middle Selective Less Selective 1995 – 1996 In-State Students Tuition Institutional Grant Unmet Need State App per FTE (N = 5695) 3584 / 1282 444 / 1338 3390 / 4028 6851 / 2490 (N = 2406) 3967 / 1408 483 / 1521 3394 / 4178 7861 / 2682 (N = 2307) 3350 / 1061 423 / 1212 3381 / 3965 6299 / 2134 (N = 979) 3164 / 1167 396 / 1103 3389 / 3779 5583 / 1639 (N = 1209) 10042 / 3990 1398 / 3287 8644 / 4988 6719 / 2518 (N = 642) 11435/ 4330 1441 / 3455 9993 / 5293 7373 / 2847 (N = 453) 8522 / 2797 1502 / 3309 7020 / 4290 6263 / 1944 (N = 113) 10581 / 14939 878 / 2249 7647 / 3728 5191 / 1251 (N = 5844) 4001 / 1756 565 / 1603 3496 / 3889 7111 / 2553 (N = 2572) 4372 / 1857 699 / 1874 3472 / 3935 8120 / 2813 (N = 2452) 3633 / 1307 471 / 1387 3445 / 3828 6455 / 2104 (N = 820) 3877 / 2248 411 / 1141 3723 / 3912 5766 / 1334 (N = 1035) 10095 / 4719 2090 / 4190 6068 / 6642 6900 / 2358 (N = 599) 11830 / 4676 2054 / 4429 6615 / 6856 7353 / 2544 (N = 362) 7693 / 3573 2011 / 3673 5184 / 6327 6422 / 2001 (N = 74) 7618 / 3658 2736 / 4510 5880 / 5951 5553 / 1091 Out-of-State Students Tuition Institutional Grant Unmet Need State App per FTE 1999 – 2000 In-State Students Tuition Institutional Grant Unmet Need State App per FTE Out-of State Students Tuition Institutional Grant Unmet Need State App per FTE 96 TABLE 11: Student Level Descriptive Statistics (continued) Sample Size Mean / Std. Deviation All Institutions More Selective Middle Selective Less Selective 2003 – 2004 In-State Students Tuition Institutional Grant Unmet Need State App per FTE (N = 7669) 4739 / 1542 761 / 2104 4077 / 4205 6138 / 2412 (N = 2880) 5013 / 1640 926 / 2456 3981 / 4331 7073 / 2749 (N = 3679) 4753 / 1483 659 / 1917 4108 / 4163 5726 / 2090 (N = 1101) 3958 / 1169 683 / 1650 4223 / 4009 5110 / 1534 (N = 1061) 12777 / 5025 2231 / 4267 8660 / 8041 6299 / 2378 (N = 527) 15641 / 4105 2241 / 4510 9799 / 8969 6797 / 2783 (N = 421) 10609 / 4479 1881 / 3908 7766 / 7138 5962 / 1868 (N = 111) 8353 / 2982 3684 / 4273 7046 / 6033 5363 / 1561 Out-of State Students Tuition Institutional Grant Unmet Need State App per FTE 97 Turning to the ordinary least squares regressions, Table 12 shows the results for the students paying in-state tuition, and Table 13 shows the results for the students paying out-of-state tuition. For both in-state and out-of-state listed tuition, the results do not reveal a consistent relationship between tuition and state appropriations, showing relationships of different size and sign for the different years. The extreme variation in the findings across years is difficult to explain and certainly suggests that researchers should be cautious when examining institution-level variables using only one year of NPSAS. 98 Table 12 Student Level Analysis: Linear Regression In-State Students at Four-Year Institutions State Appropriations per student Coefficient 1989 – 1990 List Tuition Net Tuition Institutional Grant Unmet Need B 0.051 0.054 - 0.003 0.035 Std. Dev. 0.009 0.010 0.005 0.029 t value 5.77 5.24 - 0.59 1.22 P>|t| 0.000 0.000 0.555 0.222 1992 – 1993 List Tuition Net Tuition Institutional Grant Unmet Need B - 0.039 - 0.031 - 0.007 - 0.044 Std. Dev. 0.007 0.010 0.008 0.026 t value - 5.27 - 3.09 - 0.98 - 1.68 P>|t| 0.000 0.002 0.326 0.094 1995 – 1996 List Tuition Net Tuition Institutional Grant Unmet Need B 0.037 0.025 0.012 0.038 Std. Dev. 0.006 0.009 0.008 0.021 t value 6.63 2.66 1.47 1.84 P>|t| 0.000 0.008 0.142 0.066 1999 – 2000 List Tuition Net Institutional Grant Unmet Need B - 0.006 - 0.042 0.036 - 0.004 Std. Dev. 0.009 0.013 0.009 0.019 t value - 0.64 - 3.38 3.84 - 0.19 P>|t| 0.524 0.001 0.000 0.846 2003 – 2004 List Tuition Net Institutional Grant Unmet Need B - 0.019 - 0.025 0.006 0.001 Std. Dev. 0.006 0.011 0.011 0.017 t value - 3.17 - 2.16 0.52 0.05 P>|t| 0.002 0.031 0.602 0.961 99 TABLE 13 Student Level Analysis: Linear Regression Out-of-State Students at Four Year Institutions State Appropriation Coefficients 1989 – 1990 List Tuition Net Tuition Institutional Grant Unmet Need B - 0.099 - 0.036 - 0.063 0.043 Std. Dev. 0.054 0.059 0.024 0.136 t value - 1.83 - 0.61 - 2.58 0.32 P>|t| 0.068 0.542 0.010 0.751 1992 – 1993 List Tuition Net Tuition Institutional Grant Unmet Need B 0.068 - 0.070 0.139 0.170 Std. Dev. 0.048 0.060 0.041 0.089 t value 1.41 - 1.16 3.38 1.91 P>|t| 0.159 0.245 0.001 0.057 1995 – 1996 List Tuition Net Tuition Institutional Grant Unmet Need B 0.067 0.025 0.086 - 0.047 Std. Dev. 0.045 0.009 0.044 0.073 t value 1.49 2.66 1.96 - 0.65 P>|t| 0.138 0.008 0.051 0.518 1999 – 2000 List Tuition Net Tuition Institutional Grant Unmet Need B - 0.015 - 0.110 0.121 - 0.094 Std. Dev. 0.016 0.055 0.060 0.074 t value - 0.97 - 1.99 2.00 - 1.26 P>|t| 0.331 0.047 0.046 0.209 2003 – 2004 List Tuition Net Tuition Institutional Grant Unmet Need B - 0.136 - 0.210 0.074 - 0.176 Std. Dev. 0.053 0.079 0.062 0.087 t value - 2.55 - 2.67 1.21 - 2.03 P>|t| 0.011 0.008 0.227 0.043 100 Because the net tuition paid by students falls by the amount of institutional grants they received, one might expect a more negative relationship between net tuition and state appropriations if institutions indeed use these funds for institutional aid. However, as would be anticipated by the discussion in the above paragraph, the results do not support this conclusion and yield contradictory findings from year to year as those obtained for list tuition. For in-state students the coefficients for state appropriations with net tuition as the dependent variable have magnitudes below 0.05 (except for 1989/90 where it is 0.054) that change in sign from year to year. The results are more consistent for out-of-state students, as the sign is negative for four of the five years. The results are especially strong for the last two years with coefficients of – 0.110 and – 0.210. Turning to the dependent variable for institutional grants awarded to students, one would anticipate positive coefficients for state appropriations if these funds indeed support institutional aid to students. For the in-state students, with the exception of one year the results yield low statistical significance suggesting little relationship (which is different from an inconsistent relationship that has statistical significance) between the state appropriations institutions receive and the institutional grants provided to students. The most probable explanation is that the state appropriations received by schools simply do not influence the institutional aid provided to in-state students. For the out-of-state students, with the exception of the first year, the results are again more consistent, suggesting that a $100 increase in state appropriations is associated with between a $7.40 (2003/04) and a $13.90 (1995/96) increase in institutional grants to out-of-state students. 101 In this study, the student’s unmet need is the difference between a student’s calculated expected family contribution and the costs actually incurred after financial aid not including loans. Before conducting the study, I expected this result to be more statistically significant and compelling than the analysis of tuition levels. However, the same inconsistencies and lack of statistical significance found occur with unmet need. The full regression analysis of tuition for in-state students at four-year institutions is provided in Appendix Table 4. The same regression for out-of-state students is provided in Appendix B Table 5. The results for the other independent variables, while also demonstrating considerably more volatility than the institutional level data, are more consistent across time than the results for state appropriations. The variables associated with the student characteristics of ethnicity and gender remain rather small and insignificant in terms of influencing tuition price. The general themes of higher tuition in the New England and Mid East regions also occur in the student data. Student Level Analysis – Two Year Institutions The student level regression for those attending two-year institutions is shown in Table 14. The results are for students classified as paying in-state tuition. The results reflect the same lack of consistency from year to year and demonstrate little or no relationship between the dependent variables and state appropriations The NPSAS sample populations for students paying out-of-state tuition at two-year institutions were so small that I dropped the analysis of these samples from the study. The complete student level regression of those at two-year institutions paying instate tuition is provided in Appendix Table 6. The results are similar to those found for 102 the four-year students in terms of ethnicity and gender, i.e. no particular relationship with the average tuition paid by the students. The results for governance match the negative relationship found at the institutional level. For two-year institutions, strongly regulated governing boards are associated with lower tuition levels. Though with different proportions, the theme of higher tuition in New England and the Mid East and lower tuition in the Southeast and Southwest also occurred. 103 Table 14 Student Level Analysis: Linear Regression In-State Students at Two Year Institutions State Appropriations per student Coefficient 1989 – 1990 List Tuition Net Tuition Institutional Grant Unmet Need B - 0.001 - 0.013 0.012 - 0.121 Std. Dev. 0.007 0.012 0.007 0.088 t value - 0.12 -1.10 1.60 -1.38 P>|t| 0.907 0.273 0.110 0.168 1992 – 1993 List Tuition Net Tuition Institutional Grant Unmet Need B 0.003 0.004 0.001 0.350 Std. Dev. 0.004 0.006 0.004 0.179 t value 0.80 0.62 - 0.14 1.96 P>|t| 0.423 0.536 0.887 0.053 1995 – 1996 List Tuition Net Tuition Institutional Grant Unmet Need B - 0.008 - 0.015 0.024 0.237 Std. Dev. 0.018 0.028 0.024 0.099 t value - 0.51 - 0.54 0.99 2.41 P>|t| 0.611 0.586 0.325 0.016 1999 – 2000 List Tuition Net Tuition Institutional Grant Unmet Need B 0.061 - 0.146 0.005 - 0.166 Std. Dev. 0.018 0.127 0.010 0.059 t value 3.24 - 1.15 0.47 - 2.81 P>|t| 0.001 0.254 0.638 0.005 2003 – 2004 List Tuition Net Tuition Institutional Grant Unmet Need B 0.044 0.114 - 0.070 -0.024 Std. Dev. 0.009 0.024 0.022 0.050 t value 4.49 4.81 - 3.19 - 0.48 P>|t| 0.000 0.000 0.001 0.632 104 Summary of Results This study examined the relationship between state appropriations and tuition pricing and financial aid. The analysis was conducted on institutional level data from the Integrated Postsecondary Educational Data System (IPEDS) and student level data from the National Postsecondary Student Aid Study (NPSAS) for five academic years spanning the period from 1989/90 to 2003/04. I explored the following fundamental questions. The Institutional Level Analysis 1. Given a cross section of four-year and two-year public colleges and universities, how does the list price of tuition vary as state appropriations per student in public funding changes? The institutional level analysis conducted for each year found a negative but statistically weak relationship between in-state tuition levels and the state appropriations per full time equivalent student received by the schools. Further, the relationship grows weaker over the period of the study, suggesting the higher education institutions are decoupling state appropriations from their tuition policies. For the four-year institutions, results did vary by selectivity, with the less selective schools reacting more strongly to changes in state appropriations. For out-of-state tuition, a positive relationship was found, suggesting that higher state appropriation levels may correlate with higher institutional prestige and the ability to charge higher out-of-state tuition. For the twoyear institutions, neither the in-district, in-state or out-of-state tuition levels demonstrated any relationship with state appropriations. 105 The institutional level fixed effects regression of the study is perhaps the most robust analysis performed in this study, analyzing panel data combining the information from all five years. This analysis implicitly controls for all institution-level characteristics that are constant over time, by relying purely on variation over time for individual institutions. It confirmed the negative relationship between state appropriations and in-state tuition but found a stronger correlation of about seven cents on the dollar, or that a $100 decrease in state appropriations is associated with a $7 increase in in-state tuition. It found a similar relationship for out-of-state students, which contradicts the out-of-state findings from individual year OLS regressions. The fixed effects results suggest that a drop of $100 in state appropriations per student is associated with a tuition increase of approximately $7.00, both in-state and out-of-state. Student Level Analysis 2. When looking at a cross section of college students throughout the country, how do the list tuition, net tuition, institutional grant, and their unmet need vary as the state appropriation level per student changes? The NPSAS data produced puzzling inconsistencies in the results of the single year OLS regressions, yielding coefficients for the variable relationships that varied in size and sign from year to year. These inconsistencies occurred for all four of the dependent variables. While greater volatility can be expected since the NPSAS samples consist of different students, the findings are difficult to explain. The student level analysis of the NPSAS data show a small and weakening relationship between in-state tuition and state appropriation levels when controlling for 106 the additional factors. The changing signs and significance of the coefficients over the years suggest that the level of state appropriations received by institutions does not play a role in the tuition or institutional aid across students in a given academic year. This holds for both in-state and out-of-state students. 107 CHAPTER FIVE CONCLUSION Introduction This chapter begins with a description of the key findings and a discussion of what these results are saying about what is happening in the finance of higher education, and what it all means for both institutions and students. I will address some of the implications the study may have for higher education policy makers and the leaders of higher education institutions. The chapter concludes with an exploration of interesting possibilities for future research that would build on the work performed here. The Key Findings As shown in Table 15, the cross sectional institutional level analysis using IPEDS data confirmed the literature and existing understanding that state appropriations has a suppressing effect on the in-state tuition charged at public colleges and universities. However, when looking at this relationship over the fourteen year period of the study, this relationship is declining as the share of state appropriations of total institutional budgets declines. For out-of-state tuition, the influence of state appropriations may have already ceased suppressing tuition entirely, for a positive relationship was found. Schools receiving more state support charged higher out-of-state tuition, but that relationship also weakened over time. Amidst all of the complexity of higher education finance, the cross sectional IPEDS analysis points to a rather simple conclusion. As state 108 appropriations become a smaller part of the budget of higher education institutions, so does its influence on tuition prices and quite possibly everything else. Table 15 Summary of Results – Tuition Regressions Four Year Institutions State Appropriation Coefficients In-State Analysis IPEDS NPSAS Out-of-State Analysis IPEDS NPSAS 1989 – 1990 1992 – 1993 1995 – 1996 1999 – 2000 2003 – 2004 - 0.043 - 0.031 - 0.026 - 0.015 - 0.014 0.055 0.071 0.062 0.053 0.047 Fixed Effects - 0.071 0.051 - 0.039 0.037 - 0.006 - 0.019 - 0.099 0.068 0.067 - 0.015 - 0.136 - 0.069 The student level analysis using the NPSAS data tells a story for each of the five years examined, but unfortunately the stories vary dramatically over time. For both instate and out-of-state tuition, for some years state appropriations appears to suppress tuition for some years, and in other years the opposite occurs. I had hoped to find compelling results regarding the relationship between state appropriations and institutional grants provided to students, the net tuition they paid after receiving financial aid, and the unmet need faced by students attending college. The NPSAS data did not produce results with the consistency or the magnitudes to support such findings. The most compelling analysis conducted in this study is the fixed effects regression performed on the panel data set provided by the five years of the IPEDS 109 institutional level data. Because the fixed effects regression accounts for changes over time and controls for institutional traits that remain unchanged, it examines how an individual institution’s tuition changes over time with its level of state appropriations. Interestingly, the fixed effects model produced remarkably similar results for both instate and out-of-state tuition in terms of their relationship with the state appropriations received by institutions. For each, the model suggests an association of seven cents on the dollar. A drop in state appropriations of one dollar increases both in-state and out-ofstate tuition by seven cents. In one area, all of the analyses agree. The institutional level for the individual years, the fixed effects regression using panel data, and the student level analyses all concur that no relationship exists between the level of state appropriations received and the tuition prices set at two-year institutions. The Big Picture I have examined the data for the relationship between state appropriations and tuition prices at higher education institutions in many different ways in an attempt to quantify as precisely as possible any association that may exist between the two at institutions across the country. The fixed effects analysis found that on average, a one dollar drop in state appropriations per student was associated with a seven cent increase in tuition. This result occurred for both in-state and out-of-state tuition at the four-year institutions.. The institutional level analysis of the individual years for in-state tuition at fouryear institutions also finds a negative relationship between state appropriations and 110 tuition prices, but one that is smaller than the results of the fixed effects analysis. The single year OLS regressions also yield a coefficient steadily shrinking in size over the five years, and steadily shrinking in statistical significance as well. These results are consistent with (but do not prove) the view that a stronger relationship between state appropriations and in-state tuition at four-year institutions may have existed in the past, but it is now small and declining. While the fixed effects analysis over time found a negative relationship between out-of-state tuition at four-year institutions and state appropriations, the single year OLS institutional level regressions found a positive relationship, suggesting that high state appropriations institutions charge higher out-of-state tuition. It is feasible that the schools receiving greater amounts of state appropriations have more prestige and can charge higher tuition premiums for nonresident students. Further, the student level analyses of the NPSAS data shows that institutional aid is higher (and the net tuition is lower) at the schools with high state appropriation levels. Higher state appropriations may allow four-year institutions to use institutional aid to attract out-of-state students with desirable characteristics. All of the analyses conducted suggest that no relationship exists between the state appropriations received by two-year institutions and their tuition prices. Since two-year institutions are focused on low cost programs and also subsidized by local community funding, this result is not surprising. Since community colleges are not selective, it is unlikely that they have enrollment management practices seeking to optimize net tuition revenue or other measures associated with the students who enroll. 111 Implications The continuing decline in the share of state appropriations in higher education budgets has caused universities to raise tuition and develop alternative revenue streams to meet the increasing costs of educating students. The study suggests that while over time state appropriations are a factor when considering tuition levels, they play a small role in conjunction with many other factors that include the region of the country, the politics of the state legislature, the school’s selectivity, and the demand that exists among students. State appropriations remain a critical element of university budgets, but the regressions run in this study find little relationship within a given year between the list or net tuition students experience and the appropriations provided to the institution they attend. The role of state appropriations in setting tuition levels appears to be one that occurs over time. As noted in chapter two, basic economic theory and the need to balance institutional budgets predicts a negative relationship between state appropriations and tuition. Both the fixed effects regressions and the descriptive statistics support this relationship, and in particular, that the relationship occurs over time. When isolated to a single year, however, the results suggest that variation in state appropriation levels has little or no influence on tuition levels. Further, what little influence was found for instate tuition is decreasing from year to year. Overall, the relationship found is relatively weak and may be decreasing over time. Bowen’s revenue theory suggest institutions charge what they can irrespective of state appropriations, and it is possible that tuition 112 prices are set within governance structures and other political and institutional considerations that place little emphasis on state appropriations. The single year OLS regressions suggest that the influence of state appropriations on tuition is weakening. By declining as a factor, the funds lose their desired intent of suppressing tuition and enhancing access to higher education. If the trends identified in this study continue, the role of state appropriations in subsidizing higher education will fall further, and tuition will continue to rise at rates far exceeding the rate of inflation. Students from low income backgrounds, in particular first generation students less familiar with the financial aid system and more prone to “sticker shock” are likely to view education as increasingly out of reach. The trends noted in the descriptive statistics point to the continuing stratification of institutions in terms of selectivity and prestige, with the differences between the state appropriations and tuition of the more selective schools expanding away from that of the less selective schools. As this trend continues, public schools at the top of the prestige hierarchy can become as inaccessible to lower income populations as the traditionally expensive elite private institutions. The drift away from substantial state subsidization and low tuition prices significantly increases the financial component of the decision to pursue higher education. An unintended implication of the study involves the inconsistencies found in the NPSAS data and the warning that caution should be used when using student level data to infer results regarding institutions. 113 Future Research The perplexing inconsistencies found in the NPSAS data suggest that further research is warranted to clarify the understanding of its variables and their values in a way that can apply to institutional level studies. It would be useful to see how the NPSAS tuition data for students compares to the tuition values institutions report to IPEDS for the same academic years. Even when the student populations were restricted to full time dependent students attending only one four-year institution in a year, this study found encountered highly inconsistent data. Research identifying and explaining the phenomena behind these inconsistencies could prove valuable to those using NPSAS data in their research. Another interesting study would be to conduct an analysis similar to this one for the private not-for-profit institutions replacing state appropriations received per full time student with endowment income per full time student. In particular, I would anticipate that schools with generous endowment income are in the position to discount tuition or offer institutional grants to the students meritorious enough to be admitted. I think a stronger correlation between endowment income and institutional aid exists than what this study could find between state appropriations and institutional aid. Such a study should also note the distinction between list tuition, which is unlikely to have a compelling story, and what the students actually pay, which may prove very interesting. Several of the highly endowed, prestigious elite institutions have recently adopted policies that offer extraordinary financial aid even to students with middle and upper 114 middle class backgrounds. A study analyzing this for all private non-profit higher institutions would be an interesting contribution to the literature. For those interested in deepening the understanding of the consequences of increasing stratification, future research could keep the more precise disaggregation of the selectivity categories utilized by the Barron’s taxonomy and conduct an analysis similar to this one. This may provide insight into the confusing differences found here where the more selective and less selective schools showed a negative relationship between state appropriations and tuition, while the middle selective group showed a positive one. As mentioned, the selectivity of a public four-year institution strongly influences the nature of its enrollment management context. It would also be interesting to disaggregate four-year institutions by other classifications, such as the region of the country and a measure I did not consider, the size of the institutions or their Carnegie classification. Further, this study examined the cross section of institutions and students one year at a time to explore the relationship between state appropriation levels and tuition and financial aid. I found that as one moves from one institution (or student) to others, there is very little association between the appropriation levels and the tuition prices. The fixed effects regression, which takes into account changes over time, found the opposite relationship between out-of-state tuition and state appropriations. A study that creates variables for year to year changes in state appropriations and the year to year changes in tuition pricing may find a stronger association between them. Such a study could also 115 perform the analysis with the tuition change lagging the appropriation change by one or more years. The findings of the study also point to the value of further research in the alternative revenue “other” that is growing faster than tuition. Quantifying the amounts received via fundraising, research grants, room and board, bookstores (most schools pay 75% or less of retail for textbooks and sell at full price), attractive and cash hungry student unions, and the myriad of other revenue enhancing activities schools have discovered could provide an interesting contribution to the understanding of higher education finance. Conducting these studies for the different groupings mentioned in the previous paragraph could yield valuable insights. Conclusion The United States and virtually every country will continue their commitment to provide substantial funding and resources for the education and higher education of its citizens. The higher education system will not be allowed to fail. The task at hand is more accurately viewed as a challenge of seeking optimization. What is the best higher education system and what is the best way to allocate resources for it? The question has no easy answer. Saddled with competing and compelling obligations to Medicaid, corrections, K-12 education, and other demands, state appropriations for higher education are like to continue to stagnate or decline. Sadly, the recent history in the United States has not been encouraging. The country has drifted far a field from the vision of Johnson’s Great Society. Reagan spoke of “trickle down” economics, and in some 116 respects trickle down forces have been implemented. The federal government has squeezed the states, and the squeeze has trickled down to state support for higher education. State legislatures must wrestle with their relationship with higher education. What role do states want in the college education of their residents? As appropriations continue to decline in university budgets, state relevance to the system approaches a tipping point where universities would forego (or only accept without strings) minimally significant financial support in exchange for the autonomy from state legislatures. The liberation of this autonomy is likely to be an illusion, for the benefactors filling the void are sure to voice their sentiments. Who are the likely future benefactors of higher education? If the spoken goals of the Obama administration prevail in federal education policy, students armed with generous federal financial aid will line up at higher education institutions, and their objectives are almost certain to focus on employment. The other benefactors are likely to be the future employers of those students, and the mission of higher education will be the manufacturing of skilled employees trained in the context of economic utility. Knowledge for its own sake has ironically once again become a luxury reserved for the elite and well connected few attending a small number of higher education institutions utterly inaccessible to the general population. The 1600s have emerged in the 21st Century. 117 APPENDIX A IPEDS INSTITUTIONAL LEVEL DATA SOURCES IPEDS – Institution level data is accessed from the online IPEDS STATA data sets. 2003-2004 Directory Enrollment Finance Institutional Characteristics DATA SET hd2004_data_stata.csv ef2004a_data_stata.csv f0304_f1a_data_stata.csv ic2003_ay_data_stata.csv Variables Name, Level, Control, Sector Enrollment populations State Appropriations Tuition and Fees 1999-2000 Directory Enrollment Finance Institutional Characteristics ic99_hd_data_stata.csv ef99_anr_data_stata.csv f9900_f1_data_stata.csv ip1999_ay_data_stata.csv Name, Level, Control, Sector Enrollment populations State Appropriations Tuition and Fees 1995-1996 Directory Enrollment Finance Institutional Characteristics ic9596_a_data_stata.csv ef95_anr_data_stata.csv f9596_a_data_stata.csv ic9596_b_data_stata.csv Name, Level, Control, Sector Enrollment populations State Appropriations Tuition and Fees 1992-1993 Directory Enrollment Finance Institutional Characteristics ic1992_a_data_stata.csv ef1992_a_data_stata.csv f1992_a_data_stata.csv ic1992_b_data_ stata.csv Name, Level, Control, Sector Enrollment populations State Appropriations Tuition and Fees 1989-1990 Directory Enrollment Finance Institutional Characteristics ic90hd_data_stata.csv ef1989_a_data_stata.csv f8990_a_data_stata.csv ic1989_b_data_ stata.csv Name, Level, Control, Sector Enrollment populations State Appropriations Tuition and Fees 118 APPENDIX TABLE 1 Institution Level Analysis – Four Year Institution Regression Results Dependent Variable: In-State Tuition 2004 Dollars Number of Institutions Adjusted R-Squared 1989/90 1992/93 1995/96 1999/00 2003/04 453 0.4397 468 0.5700 460 0.6018 463 0.5749 432 0.4446 - 0.015 / 0.016 255 / 128 60 / 113 1397 / 178 773 / 137 259 / 120 - 55 / 137 1924 / 198 1865 / 151 1427 / 158 711 / 176 244 / 153 - 155 /169 245 / 207 2441 / 203 - 0.014 / 0.025 362 / 223 - 85 / 168 2143 / 276 1138 / 197 434 / 174 62 / 201 1560 / 339 1354 / 301 1235 / 309 539 / 315 - 181 / 287 - 754 / 309 - 506 / 364 3555 / 343 Coefficient / Standard Error State Appropriations Govern1 Govern2 Most Selective More Selective Middle Selective Less Selective New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Constant / Intercept - 0.043 / 0.013 387 / 111 - 102 / 99 1176 / 159 498 / 120 238 / 124 0 / 119 1291 / 184 1164 / 132 1280 / 146 597 / 162 597 / 141 - 151 / 157 507 / 186 1741 / 186 - 0.031 / 0.016 - 0.026 / 0.017 590 / 117 677 / 124 274 / 104 453 / 109 1179 / 163 1349 / 171 554 / 125 651 / 130 203 / 109 195 / 114 - 105 / 124 - 107 / 130 2151 / 118 2088 / 191 1445 / 139 1492 / 145 1141 / 149 1152 / 153 568 / 167 525 / 172 342 / 147 179 / 149 - 533 / 161 - 671 / 167 307 / 194 169 / 200 2074 / 192 2389 / 194 = 119 APPENDIX TABLE 2 Institution Level Analysis – Four Year Institution Regression Results Dependent Variable: Out-of-State Tuition 2004 Dollars Number of Institutions Adjusted R-Squared 1989/90 1992/93 1995/96 1999/00 2003/04 453 0.4560 468 0.4940 460 0.4787 463 0.4463 432 0.3650 0.062 / 0.045 2517 / 338 1529 / 297 4071 / 465 2475 / 354 824 / 309 507 / 354 2190 / 518 - 977 / 394 537 / 415 - 1489 / 466 - 1465 / 404 - 2023 / 453 436 / 541 7177 / 528 0.070 / 0.041 1515 / 342 964 / 302 4525 / 474 2851 / 365 934 / 322 454 / 366 1843 / 530 - 1396 / 404 561 /424 - 1757 / 474 - 1261 / 413 - 2228 / 451 265 / 555 7952 / 551 0.047 / 0.060 968 / 544 175 / 409 6712 / 674 4230 / 480 1697 / 424 1083 / 490 951 / 827 - 1298 / 733 1210 / 752 - 1598 / 767 - 874 / 699 - 2656 / 752 - 506 / 364 9290 / 836 Coefficient / Standard Error State Appropriations Govern1 Govern2 Most Selective More Selective Middle Selective Less Selective New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Constant / Intercept 0.055 / 0.031 1689 / 262 - 102 / 99 3361 / 373 2006 / 283 731 / 246 511 / 280 1293 / 432 - 1064 / 311 998 / 342 - 1354 / 382 - 784 / 331 - 1139 / 369 34 / 437 5114 / 439 0.071 / 0.042 2456 / 308 1010 / 273 3955 / 427 2406 / 328 885 / 286 469 / 326 2911 / 495 - 1003 / 364 781 / 393 - 1197 / 439 - 830 / 385 - 1723 / 423 466 / 509 5840 / 505 = 120 APPENDIX TABLE 3 Institution Level Analysis – Two Year Institution Regression Results Dependent Variable: In-State Tuition 2004 Dollars Number of Institutions Adjusted R-Squared 1989/90 1992/93 1995/96 1999/00 2003/04 858 0.6310 864 0.6127 797 0.6296 941 0.6115 900 0.5254 - 0.001 / 0.002 - 847 / 87 - 3 / 77 1002 / 144 1715 / 106 2035 / 100 644 / 144 - 250 / 96 - 133 /97 427 / 155 1421 / 90 0.001 / 0.003 - 1119 / 108 - 405 / 97 828 / 179 2113 / 128 1898 / 123 839 / 140 - 246 / 119 285 / 125 113 / 186 1970 / 111 Coefficient / Standard Error State Appropriations Govern1 Govern2 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Constant / Intercept - 0.053 / 0.011 - 620 / 82 - 347 / 73 903 / 129 1901 / 98 2023 / 94 724 / 110 105 / 93 428 / 157 453 / 186 1212 / 96 - 0.008 / 0.004 - 0.034 / 0.010 - 562 / 83 - 516 / 81 - 264 / 74 - 215 / 69 1360 / 134 1305 / 127 2243 / 98 2243 / 93 1845 / 95 1468 / 94 786 / 111 631 / 103 80 / 92 - 71 / 88 380 / 96 302 / 90 466 / 147 397 / 133 1164 / 87 1500 / 91 121 APPENDIX TABLE 4 Student Level Analysis - In-State Tuition Regression Students Attending Four-Year Institutions 2004 Dollars 1989/90 1992/93 1995/96 Student Observations Adjusted R-squared 3352 0.1877 5433 0.3215 5695 0.5320 Coefficient / Standard Error State Appropriations Expected Family Contribution Cumulative Grade Point Average Strongly Regulated Moderately Regulated Most Selective More Selective Middle Selective Less Selective New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Female White African American Asian Hispanic Constant / Intercept 0.051 / 0.008 0.000 / 0.001 - 0.187 / 2.96 - 226 / 66 - 167 / 74 514 / 167 - 408 / 152 - 367 / 149 - 491 / 157 1091 / 105 552 / 79 621 / 75 150 / 89 48 / 80 - 133 / 85 832 / 94 - 40 / 40 12 / 272 472 / 283 285 / 288 183 / 288 1223 / 324 - 0.039 / 0.007 0.002 / 0.001 1.385 / 0.260 889 / 65 79 / 55 622 / 82 325 / 75 - 86 / 71 - 353 / 86 2541 / 115 1516 / 72 1349 / 78 922 / 83 572 / 76 - 561 / 78 - 923 / 107 - 38 / 120 - 448 / 93 - 262 / 109 - 38 / 120 - 613 / 115 2388 / 157 - 0.037 / 0.005 0.004 / 0.001 1.498 / 0.189 980 / 42 716 / 42 788 / 73 390 / 66 183 / 64 - 183 / 69 3033 / 86 1749 / 50 1307 / 50 690 / 66 153 / 54 - 473 / 62 650 / 74 - 53 / 24 11 / 112 99 / 120 156 / 119 - 204 / 122 1598 / 149 122 APPENDIX TABLE 4 (continued) Student Level Analysis – In-State Tuition Regression Students Attending Four-Year Institutions 2004 Dollars 1999/00 2003/04 Student Observations Adjusted R-squared 5844 0.2848 7669 0.4349 Coefficient / Standard Error State Appropriations Expected Family Contribution Cumulative Grade Point Average Strongly Regulated Moderately Regulated Most Selective More Selective Middle Selective Less Selective New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Female White African American Asian Hispanic Constant / Intercept - 0.006 / .009 0.005 / .002 0.812 / 0.316 494 / 74 222 / 65 902 / 110 209 / 100 - 237 / 96 178 / 111 2396 / 127 1647 / 83 913 / 86 449 / 101 - 150 / 84 - 790 / 91 - 129 / 130 74 / 40 257 / 108 135 / 131 114 / 131 -42 / 131 2869 / 185 - 0.019 / .006 0.004 / .001 0.643 / 0.211 -252 / 59 -122 / 41 1112 / 76 225 / 64 95 / 62 -474 / 76 1571 / 88 1780 / 71 1279 / 73 508 / 81 -468 / 69 -492 / 76 -605 / 86 -5.21 / 27 76 / 66 323 / 80 41 / 91 47 / 84 3953 / 129 123 APPENDIX TABLE 5 Student Level Analysis – Out-of-State Tuition Regression Students Attending Four-Year Institutions 2004 Dollars 1989/90 1992/93 1995/96 Student Observations Adjusted R-squared 395 0.4003 1233 0.3191 1209 0.3714 Coefficient / Standard Error State Appropriations Expected Family Contribution Cumulative Grade Point Average Strongly Regulated Moderately Regulated Most Selective More Selective Selective Less Selective New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Female White African American Asian Hispanic Constant / Intercept - 0.104 / 0.050 0.002 / 0.005 23.36 / 16.75 946 / 412 236 / 431 6972 / 988 4824 / 881 3671 / 899 3416 / 960 1334 / 541 - 860 / 492 699 / 423 441 / 514 444 / 451 340 / 523 4179 / 471 - 177 / 221 384 / 1596 2065 / 1637 855 / 1650 - 298 / 1780 - 1549 / 1899 0.068 / 0.048 0.010 / 0.006 - 1.27 / 1.72 2216 / 425 708 / 354 4068 / 571 2583 / 513 949 / 494 - 792 / 609 7592 / 706 2913 / 631 1306 / 587 292 / 607 1243 / 565 57 / 597 2931 / 711 15 / 202 - 432 / 485 - 569 / 597 1269 / 617 - 559 / 737 3733 / 1026 0.067 / 0.045 0.026 / 0.006 0.209 / 1.49 1672 / 403 865 / 340 1584 / 520 1208 / 486 - 837 / 458 - 974 / 566 6674 / 672 1521 / 583 566 / 542 - 143 / 620 44 / 559 - 2102 / 586 518 / 663 22 / 185 - 804 / 603 - 193 / 687 174 / 673 - 1249 / 789 8627 / 1027 124 APPENDIX TABLE 5 (continued) Student Level Analysis – Out-of-State Tuition Regression Students Attending Four-Year Institutions 2004 Dollars 1999/00 2003/04 Student Observations Adjusted R-squared 1035 0.3679 1061 0.4892 Coefficient / Standard Error State Appropriations Expected Family Contribution Cumulative Grade Point Average Strongly Regulated Moderately Regulated Most Selective More Selective Selective Less Selective New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Female White African American Asian Hispanic Constant / Intercept - 0.109 / 0.055 0.048 / 0.009 0.089 / 1.96 654 / 502 368 / 401 4802 / 833 2551 / 775 - 465 / 775 358 / 919 5674 / 663 806 / 611 60 / 571 - 980 / 613 6.35 / 543 - 1399 / 607 1638 / 713 153 / 243 301 / 627 - 859 / 723 1633 / 759 71 / 880 7556 / 1211 - 0.136 / .053 0.015 / .007 3.647 / 1.863 - 140 / 815 - 565 / 374 9925 / 717 8816 / 605 4373 / 609 2936 / 773 3566 / 668 3190 / 637 2986 / 616 - 2127 / 602 471 / 556 -1338 / 637 2241 / 673 - 3 / 233 - 1507 / 633 634 / 698 - 488 / 745 - 218 / 837 6075 / 1209 125 APPENDIX TABLE 6 Student Level Analysis - In-State Tuition Regression Students Attending Two-Year Institutions 1989/90 1992/93 1995/96 Student Observations Adjusted R-squared 267 0.4407 398 0.5888 499 0.6479 Coefficient / Standard Error State Appropriations Exp. Family Contribution Cum.Grade Point Average Strongly Regulated Moderately Regulated New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Female White African American Asian Hispanic Constant / Intercept - 0.001 / 0.007 0.002 / 0.002 - 0.916 / 2.66 - 143 / 73 158 / 65 563 / 115 575 / 71 288 / 85 219 / 99 40 / 74 - 41 / 103 832 / 94 - 1.8 / 40 - 81 / 312 - 167 / 320 18 / 324 - 116 / 316 586 / 323 0.003 / 0.004 0.005 / 0.002 0.524 / 0.424 - 468 / 123 228 / 104 1570 / 252 1450 / 130 1410 / 137 1593 / 169 450 / 140 - 14 / 160 954 / 232 1.4 / 61 - 19 / 230 - 6.8 / 273 100 / 264 85 / 257 575 / 278 0.008 / 0.018 - 0.002 / 0.003 0.338 / 0.327 - 983 / 77 252 / 63 944 / 1686 964 / 970 710 / 97 63 / 110 - 363 / 96 - 570 / 99 223 / 194 100 / 45 20 / 145 - 14 / 157 - 182 / 182 - 85 / 159 1584 / 182 126 APPENDIX TABLE 6 (continued) Student Level Analysis - In-State Tuition Regression Students Attending Two-Year Institutions 1999/00 2003/04 Student Observations Adjusted R-squared 693 0.5096 3156 0.5829 Coefficient / Standard Error State Appropriations Expected Family Contribution Cumulative Grade Point Average Strongly Regulated Moderately Regulated New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Female White African American Asian Hispanic Constant / Intercept 0.062 / 0.018 - 0.002 / .003 0.464 / 0.388 - 1033 / 112 - 140 / 96 380 / 255 1720 / 119 612 / 128 558 / 162 - 301 / 129 - 82 / 132 237 / 201 0.865 / 59 239 / 129 21 / 159 428 / 195 95 / 151 1034 / 199 0.044 / 0.009 0.003 / 0.001 0.416 / 0.139 -820 / 36 125 / 35 699 / 86 1298 / 41 585 / 43 883 / 52 225 / 47 -457 / 50 41 / 158 - 40 / 20 - 39 / 52 - 239 / 80 - 130 / 70 - 132 / 84 1346 / 79 127 Appendix Table 7 Descriptive Statistics – IPEDS Institutions 2003 – 2004 Four Year Institutions (N = 432) Variable Mean Std. Dev. Min Max In-State Tuition/Fee Out-of-State Tuition/Fee State App / Student 4443.17 11,211.44 5594.72 1494.94 3406.94 2322.65 2032 4254 417.39 13,868 29,064 15,949.48 Weakly Regulated Moderately Regulated Strongly Regulated 0.081 0.194 0.725 0.273 0.396 0.447 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.058 0.206 0.433 0.171 0.123 0.234 0.405 0.496 0.377 0.328 0 0 0 0 0 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Far West 0.074 0.141 0.157 0.111 0.313 0.106 0.053 0.045 0.262 0.349 0.365 0.315 0.464 0.309 0.225 0.197 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 128 Appendix Table 7 (continued) Descriptive Statistics – IPEDS Institutions 2003 – 2004 Two Year Institutions (N = 900) Variable In-District Tuition In-State Tuition Out-of-State Tuition State App. / Student Mean 1924.71 2319.62 5287.20 4481.37 Std. Dev. 883.87 1293.37 2483.59 10691.73 Min 48 48 179 0.01 Max 7294 7391 17,630 24,078 Weakly Regulated Moderately Regulated Strongly Regulated 0.148 0.176 0.677 0.355 0.381 0.468 0 0 0 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Far West 0.044 0.099 0.150 0.106 0.309 0.112 0.039 0.141 0.206 0.299 0.357 0.307 0.462 0.316 0.193 0.340 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 129 Appendix Table 7 (continued) Descriptive Statistics – IPEDS Institutions 1999 – 2000 Four Year Institutions (N = 463) Variable Mean Std. Dev. Min Max In-State Tuition Out-of-State Tuition State App. / Student 3483.28 9463.50 6581.12 1154.05 2682.98 2448.23 1611.33 2934.68 1543.58 8841.41 21,719.92 19,160.98 Weakly Regulated Moderately Regulated Strongly Regulated 0.149 0.179 0.672 0.357 0.384 0.470 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.069 0.205 0.434 0.177 0.112 0.254 0.404 0.496 0.382 0.316 0 0 0 0 0 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Far West 0.067 0.136 0.147 0.104 0.287 0.099 0.058 0.102 0.250 0.343 0.354 0.305 0.453 0.299 0.235 0.299 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 130 Appendix Table 7 (continued) Descriptive Statistics – IPEDS Institutions 1999 – 2000 Two Year Institutions (N = 941) Variable Mean Std. Dev. Min Max In-State Tuition Out-of-State Tuition State App. / Student 1768.96 4667.94 5297.11 1148.12 2542.20 9493.90 62.65 137.39 148.71 7175.13 19,740.38 27,315.80 Weakly Regulated Moderately Regulated Strongly Regulated 0.143 0.173 0.683 0.351 0.379 0.465 0 0 0 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Far West 0.043 0.087 0.145 0.099 0.330 0.125 0.034 0.137 0.202 0.282 0.352 0.299 0.470 0.331 0.181 0.338 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 131 Appendix Table 7 (continued) Descriptive Statistics – IPEDS Institutions 1995 – 1996 Four Year Institutions (N = 460) Variable In-State Tuition Out-of-State Tuition State App. / Student Mean 3278.29 8759.35 5923.47 Std. Dev. 1147.41 2814.10 2236.98 Min 1596.06 2708 661.47 Max 8322.40 22,000.75 17,071.23 Weakly Regulated Moderately Regulated Strongly Regulated 0.143 0.183 0.674 0.351 0.387 0.469 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.067 0.202 0.437 0.172 0.115 0.251 0.402 0.497 0.376 0.320 0 0 0 0 0 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Far West 0.070 0.141 0.146 0.102 0.289 0.094 0.059 0.099 0.255 0.349 0.353 0.303 0.454 0.291 0.235 0.297 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 132 Appendix Table 7 (continued) Descriptive Statistics – IPEDS Institutions 1995 – 1996 Two Year Institutions (N = 797) Variable Mean Std. Dev. Min Max In District Tuition In-State Tuition Out-of-State Tuition State App. / Student 1590.63 1810.70 4597.46 4437.39 802.54 1014.04 2274.73 2366.53 187.91 187.91 722.74 12.66 5173.65 5432.63 15,723.30 28,887.29 Weakly Regulated Moderately Regulated Strongly Regulated 0.132 0.208 0.660 0.338 0.406 0.474 0 0 0 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Far West 0.050 0.108 0.118 0.103 0.325 0.123 0.043 0.130 0.218 0.310 0.323 0.304 0.469 0.329 0.202 0.326 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 133 Appendix Table 7 (continued) Descriptive Statistics – IPEDS Institutions 1992 – 1993 Four Year Institutions (N = 468) Variable Mean Std. Dev. Min Max In-State Tuition Out-of-State Tuition State App. / Student 2934.75 7689.33 5798.13 1058.59 2564.56 2408.77 1151.70 2158.46 595.03 8051.47 19,281.22 18,514.47 Weakly Regulated Moderately Regulated Strongly Regulated 0.150 0.179 0.671 0.357 0.384 0.470 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.068 0.205 0.432 0.175 0.111 0.253 0.404 0.496 0.381 0.315 0 0 0 0 0 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Far West 0.068 0.145 0.141 0.103 0.288 0.096 0.058 0.101 0.253 0.353 0.348 0.304 0 .454 0.295 0.233 0.289 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 134 Appendix Table 7 (continued) Descriptive Statistics – IPEDS Institutions 1992 – 1993 Two Year Institutions (N = 864) Variable Mean Std. Dev. Min Max In-District Tuition In-State Tuition Out-of-State Tuition State App. / Student 1437.70 1751.86 4431.30 3862.16 784.40 1088.34 2080.41 4921.34 104.46 104.46 143.64 52.10 6254.709 6386.593 15,205.86 13,744.10 Weakly Regulated Moderately Regulated Strongly Regulated 0.153 0.189 0.659 0.360 0.391 0.474 0 0 0 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Far West 0.050 0.105 0.156 0.097 0.289 0.123 0.037 0.143 0.218 0.307 0.363 0.296 0.454 0.328 0.189 0.342 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 135 Appendix Table 7 (continued) Descriptive Statistics – IPEDS Institutions 1989 – 1990 Four Year Institutions (N = 453) Variable Mean Std. Dev. Min Max In-State Tuition Out-of-State Tuition State App. / Student 2455.82 6461.59 6726.88 883.22 2110.15 2859.50 632.98 1787.58 846.62 5986.91 17,562.19 22,704.71 Weakly Regulated Moderately Regulated Strongly Regulated 0.152 0.181 0.667 0.360 0.385 0.472 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.068 0.196 0.433 0.181 0.115 0.253 0.398 0.496 0.385 0.319 0 0 0 0 0 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Far West 0.062 0.152 0.146 0.104 0.276 0.099 0.060 0.101 0.241 0.360 0.353 0.306 0.447 0.299 0.237 0.297 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 136 Appendix Table 7 (continued) Descriptive Statistics – IPEDS Institutions 1989 – 1990 Two Year Institutions (N = 858) Variable In-District Tuition In-State Tuition Out-of-State Tuition State App. / Student Mean 1221.18 1545.97 3842.89 4456.01 Std. Dev. 721.22 1096.44 1881.91 2335.29 Min 73.26 73.26 146.52 76.72 Max 5778.84 5975.19 13,902.06 23,646.96 Weakly Regulated Moderately Regulated Strongly Regulated 0.161 0.186 0.653 0.368 0.390 0.476 0 0 0 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Far West 0.055 0.106 0.157 0.094 0.283 0.119 0.037 0.149 0.228 0.308 0.364 0.293 0.451 0.324 0.190 0.348 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 137 Appendix Table 8 Descriptive Statistics 2003 – 2004 Four Year In-State Students (N = 7669) Variable Mean Std. Dev Min Max Tuition Net Tuition EFC Inst. Grant Unmet Need State App / Student 4738.64 3977.71 12,409.05 760.94 3602.85 6137.98 1542.32 2407.55 14,630.83 2104.18 4948.81 2412.24 1224 -16646 0 0 -35,082 1762.76 11,445 11,295 99,000 21,883 24,700 15,852.95 GPA Female White African American Asian Hispanic Other 293.54 0.553 0.772 0.084 0.041 0.060 0.033 65.02 0.498 0.421 0.287 0.201 0.238 0.120 0 0 0 0 0 0 0 400 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Far West 0.051 0.158 0.213 0.075 0.277 0.107 0.072 0.048 0.220 0.365 0.409 0.263 0.448 0.309 0.258 0.214 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.067 0.206 0.727 0.250 0.404 0.445 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.089 0.280 0.493 0.074 0.063 0.285 0.449 0.500 0.262 0.243 0 0 0 0 0 1 1 1 1 1 138 Appendix Table 8 (continued) Descriptive Statistics 2003 – 2004 Four Year Out-of-State Students (N = 1061) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / Student 12,777.01 10,545.88 14,997.73 2231.13 8003.92 6299.31 5024.84 6429.683 16,906.26 4266.88 9065.11 2378.00 3471 -12,386 0 0 - 22,500 1762.76 26,924 26,924 99,000 22,191 33,346 15,852.95 GPA Female White African American Asian Hispanic Other 297.79 0.475 0.725 0.131 0.072 0.039 0.033 61.88 0.500 0.447 0.338 0.259 0.193 0.103 64 0 0 0 0 0 0 400 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Far West 0.082 0.102 0.146 0.128 0.282 0.084 0.121 0.054 0.275 0.303 0.353 0.334 0.450 0.278 0.326 0.226 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.022 0.126 0.853 0.145 0.332 0.355 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.089 0.386 0.425 0.050 0.049 0.285 0.487 0.495 0.218 0.217 0 0 0 0 0 1 1 1 1 1 139 Appendix Table 8 (continued) Descriptive Statistics 2003 – 2004 Two Year In-State Students (N = 3156) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 1816.37 1495.72 10,022.28 320.65 2680.89 2991.71 861.65 1604.53 14,477.88 1375.82 3583.69 1320.00 378 - 18020 0 0 - 20842 164.00 5192 5079 99,000 20,842 18,163 8652.61 GPA Female White African American Asian Hispanic Other 283.17 0.513 0.679 0.125 0.042 0.114 0.040 73.53 0.500 0.467 0.331 0.202 0.318 0.131 0 0 0 0 0 0 0 400 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Far West 0.021 0.176 0.183 0.106 0.260 0.107 0.005 0.141 0.143 0.381 0.387 0.308 0.439 0.310 0.067 0.348 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.221 0.170 0.608 0.415 0.376 0.488 0 0 0 1 1 1 140 Appendix Table 8 (continued) Descriptive Statistics – NPSAS Students 1999 – 2000 Four Year In-State Students (N = 5863) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 4001.33 3435.862 12,379.84 565.47 3496.25 7110.89 1755.67 2283.93 13,056.70 1602.81 3888.81 2553.23 48.36 -14,428.29 0 0 0 1789.543 18,254.36 18,254.36 109,074.40 19,256.77 29,236.87 18,449.29 GPA Female White African American Asian Hispanic Other 288.24 0.545 0.750 0.092 0.061 0.062 0.034 63.67 0.498 0.433 0.289 0.239 0.242 0.176 0 0 0 0 0 0 0 400 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain 0.039 0.113 0.192 0.089 0.270 0.102 0.039 0.193 0.317 0.394 0.285 0.444 0.302 0.194 0 0 0 0 0 0 0 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.190 0.198 0.612 0.392 0.399 0.487 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.148 0.304 0.408 0.092 0.048 0.356 0.460 0.491 0.290 0.213 0 0 0 0 0 1 1 1 1 1 141 Appendix Table 8 (continued) Descriptive Statistics – NPSAS Students 1999 – 2000 Four Year Out-of-State Students (N = 1039) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 10095.41 8005.90 14,509.57 2089.51 6068.22 6899.52 4719.36 5911.59 13,508.54 4190.32 6642.26 2357.53 311.05 -13,155.49 0 0 0 1789.54 25,928.49 21,138.48 84,507.75 26,868.25 54,956.53 18,449.29 GPA Female White African American Asian Hispanic Other 291.24 0.520 0.752 0.108 0.067 0.035 0.037 61.52 0.500 0.432 0.311 0.249 0.184 0.179 62 0 0 0 0 0 0 400 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Far West 0.083 0.127 0.191 0.104 0.266 0.097 0.061 0.276 0.333 0.393 0.305 0.442 0.296 0.240 0 0 0 0 0 0 0 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.133 0.117 0.750 0.340 0.321 0.433 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.150 0.432 0.344 0.048 0.026 0.357 0.496 0.475 0.215 0.158 0 0 0 0 0 1 1 1 1 1 142 Appendix Table 8 (continued) Descriptive Statistics – NPSAS Students 1999 – 2000 Two Year In-State Students (N = 693) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 1689.83 1541.30 10,044.03 148.53 2388.89 3969.11 1090.62 1190.88 10,611.92 426.26 2958.26 1852.59 20.88 -1428.87 0 0 0 384.76 8988.69 8988.69 68,484.63 3028.11 13,952.36 12,039.71 GPA Female White African American Asian Hispanic Other 270.09 0.520 0.701 0.098 0.038 0.105 0.058 76.81 0.500 0.458 0.297 0.191 0.307 0.108 0 0 0 0 0 0 0 400 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain 0.017 0.166 0.175 0.063 0.242 0.109 0.031 0.128 0.372 0.380 0.243 0.428 0.311 0.174 0 0 0 0 0 0 0 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.212 0.215 0.573 0.409 0.411 0.495 0 0 0 1 1 1 143 Appendix Table 8 (continued) Descriptive Statistics – NPSAS Students 1995 – 1996 Four Year In-State Students (N = 5533) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 3583.60 3139.38 10,584.41 444.22 3389.82 6851.25 1281.77 1772.23 12,324.46 1337.56 4028.41 2490.19 327.64 -13,014.22 0 0 -16,683.35 1551.85 12,720.30 12,720.30 114,434.50 16,081.06 22,886.91 13,426.33 GPA Female White African American Asian Hispanic Other 278.38 0.542 0.781 0.080 0.074 0.054 0.011 63.49 0.498 0.414 0.271 0.262 0.226 0.075 4 0 0 0 0 0 0 400 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Far West 0.029 0.123 0.214 0.077 0.266 0.084 0.052 0.154 0.168 0.329 0.410 0.267 0.442 0.277 0.223 0.361 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.244 0.147 0.610 0.429 0.354 0.488 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.122 0.306 0.410 0.123 0.039 0.327 0.461 0.492 0.328 0.194 0 0 0 0 0 1 1 1 1 1 144 Appendix Table 8 (continued) Descriptive Statistics – NPSAS Students 1995 – 1996 Four Year Out-of-State Students (N = 1161) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 10042.34 8644.19 14,203.44 1398.15 5882.15 6719.33 3989.99 4988.13 15,024.16 3287.36 6723.89 2518.01 1911.66 -20,537.99 0 0 -15,057.17 2574.19 31,480.34 31,480.34 114,434.50 31,061.14 39,589.53 13,426.33 GPA Female White African American Asian Hispanic Other 280.95 0.490 0.771 0.078 0.093 0.034 0.025 63.41 0.500 0.421 0.268 0.291 0.180 0.132 25 0 0 0 0 0 0 400 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Far West 0.071 0.093 0.175 0.102 0.322 0.132 0.062 0.042 0.257 0.291 0.381 0.303 0.468 0.339 0.240 0.200 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.146 0.146 0.707 0.353 0.353 0.455 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.212 0.310 0.363 0.067 0.048 0.409 0.462 0.481 0.249 0.214 0 0 0 0 0 1 1 1 1 1 1 145 Appendix Table 8 (continued) Descriptive Statistics – NPSAS Students 1995 – 1996 Two Year In-State Students (N = 489) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 1677.18 1457.29 7966.49 219.88 2509.36 3591.99 796.33 945.39 8326.47 669.53 3289.73 1492.91 80.71 -3376.42 0 0 -4741.20 692.14 4599.06 3922.09 69,858.06 4818.30 14,456.09 7838.499 GPA Female White African American Asian Hispanic Other 264.90 0.472 0.750 0.115 0.033 0.078 0.025 68.89 0.500 0.434 0.319 0.180 0.268 0.138 43 0 0 0 0 0 0 400 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountains Far West 0.023 0.148 0.119 0.096 0.338 0.139 0.015 0.123 0.150 0.355 0.324 0.295 0.473 0.346 0.123 0.329 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.170 0.199 0.632 0.376 0.399 0.483 0 0 0 1 1 1 146 Appendix Table 8 (continued) Descriptive Statistics – NPSAS Students 1992 – 1993 Four Year In-State Students (N = 5433) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 3126.53 2767.51 11,986.69 359.02 5394.53 6536.62 1361.503 1731.696 14,129.09 1180.055 2678.402 2543.673 261.16 - 9474.77 0 0 0 1574.616 14,630.01 14,630.01 117,520.6 18,254.87 16,647.45 15,356.11 GPA Female White African America Asian Hispanic Other 276.21 0.512 0.810 0.069 0.043 0.050 0.028 61.22 0 .500 0.392 0.254 0.203 0.218 0.146 7 0 0 0 0 0 0 400 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Mountain Far West 0.032 0.110 0.173 0.123 0.298 0.108 0.041 0.116 0.175 0.312 0.378 0.329 0.457 0.310 0.198 0.320 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Highly Regulated 0.151 0.169 0.678 0.358 0.375 0.467 0 0 0 1 1 1 More Selective Most Selective Middle Selective Less Selective Non Selective 0.149 0.316 0.398 0.081 0.055 0.357 0.465 0.490 0.273 0.228 0 0 0 0 0 1 1 1 1 1 147 Appendix Table 8 (continued) Descriptive Statistics – NPSAS Students 1992 – 1993 Four Year Out-of-State Students (N = 1233) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 7339.24 6214.39 14,192.56 1124.85 8091.14 6192.68 4218.65 4964.99 16,925.3 2975.71 4386.17 2414.69 407.40 -13,979.73 0 0 0 1591.75 31,338.83 30,685.94 117,520.6 20,761.98 34,420.48 15,356.11 GPA Female White African American Asian Hispanic Other 283.05 0.514 0.762 0.095 0.067 0.031 0.045 60.53 0.500 0.426 0.293 0.250 0.174 0.205 100 0 0 0 0 0 0 400 1 1 1 1 1 1 New England MidEast Great Lakes Plains Southeast Southwest Rocky Far West 0.060 0.075 0.149 0.137 0.387 0.092 0.055 0.044 0.238 0.264 0.356 0.344 0.487 0.289 0.229 0.205 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.096 0.112 0.792 0.295 0.315 0.406 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.143 0.384 0.355 0.064 0.055 0.351 0.487 0.479 0.244 0.227 0 0 0 0 0 1 1 1 1 1 148 Appendix Table 8 (continued) Descriptive Statistics – NPSAS Students 1992 – 1993 Two Year In-State Students (N = 398) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 1419.67 1286.59 10,573.95 133.08 4739.52 3744.52 931.17 1088.33 14,022.53 559.16 2634.49 6984.52 78.35 -3879.49 141.02 0 0 758.15 5464.71 5464.71 117,520.60 5223.14 12,988.64 137,369.90 GPA Female White African American Asian Hispanic Other 265.50 0.490 0.808 0.047 0.058 0.067 0.019 73.14 0.502 0.394 0.213 0.235 0.250 0.108 7 0 0 0 0 0 0 400 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Far West 0.020 0.134 0.163 0.078 0.276 0.082 0.025 0.222 0.140 0.341 0.370 0.269 0.448 0.275 0.155 0.416 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.248 0.145 0.607 0.433 0.353 0.489 0 0 0 1 1 1 149 Appendix Table 8 (continued) Descriptive Statistics – NPSAS Students 1989 – 1990 Four Year In-State Students (N = 3352) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 1631.16 1483.53 13,669.98 147.6312 1737.48 7160.34 1274.29 1448.20 16,368.16 707.34 4209.32 3015.57 397.08 -15,028.81 1025.66 0 -17099.18 1729.20 36,630.63 36,630.63 144,431.6 17,099.18 41,397 19,908.83 GPA Female White African American Asian Hispanic Other 26.80 0.518 0.845 0.062 0.046 0.042 0.005 6.84 0.500 0.362 0.242 0.210 0.200 0.073 0 0 0 0 0 0 0 40 1 1 1 1 1 1 Mid East Great Lakes Plains Southeast Southwest Rocky Far West 0.148 0.221 0.111 0.151 0.113 0.105 0.132 0.356 0.415 0.314 0.358 0.317 0.306 0.339 0 0 0 0 0 0 0 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.218 0.193 0.590 0.413 0.395 0.492 0 0 0 1 1 1 Most Selective More Selective Middle Selective Less Selective Non Selective 0.102 0.268 0.501 0.110 0.019 0.302 0.443 0.500 0.313 0.138 0 0 0 0 0 1 1 1 1 1 150 Appendix Table 8 (continued) Descriptive Statistics – NPSAS Students 1989 – 1990 Four Year Out-of-State Students (N = 395) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 4291.11 4079.00 19,165.27 212.12 2483.70 7114.31 2717.05 2868.89 22,295.17 980.55 6289.78 2490.57 397.076 -5515.11 1025.66 0 -14,835.4 2745.64 16,996.61 16,996.61 136,431.50 13,157.72 31,997.59 14,946.56 GPA Female White African American Asian Hispanic Other 26.99 0.460 0.809 0.095 0.073 0.019 0.005 6.73 0.499 0.394 0.294 0.261 0.136 0.068 0 0 0 0 0 0 0 40 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Far West 0.091 0.153 0.205 0.085 0.242 0.089 0.120 0.043 0.288 0.361 0.404 0.279 0.429 0.285 0.325 0.203 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.192 0.103 0.705 0.394 0.304 0.456 0 0 0 1 1 1 Most selective More selective Middle selective Less selective Non Selective 0.120 0.361 0.392 0.110 0.017 0.326 0.481 0.489 0.314 0.129 0 0 0 0 0 1 1 1 1 1 151 Appendix Table 8 (continued) Descriptive Statistics – NPSAS Students 1989 – 1990 Two Year In-State Students (N = 269) Variable Mean Std. Dev. Min Max Tuition Net Tuition EFC Institutional Grant Unmet Need State App. / student 692.75 626.84 9092.49 65.90 1426.71 4113.41 413.12 491.45 10,435.99 267.367 3836.84 2974.57 109.89 -1002.21 1025.66 0 -5263.09 453.32 2530.44 2530.44 84,511.26 1667.43 35,650.39 23,634.90 GPA Female White African American Asian Hispanic Other 25.88 0.571 0.816 0.058 0.057 0.065 0.004 7.50 0.496 0.388 0.234 0.232 0.248 0.064 0 0 0 0 0 0 0 40 1 1 1 1 1 1 New England Mid East Great Lakes Plains Southeast Southwest Rocky Far West 0.043 0.194 0.136 0.075 0.281 0.054 0.063 0.163 0.203 0.397 0.344 0.263 0.450 0.227 0.244 0.370 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Weakly Regulated Moderately Regulated Strongly Regulated 0.159 0.220 0.622 0.366 0.415 0.486 0 0 0 1 1 1 152 REFERENCES Archibald, Robert B. & Feldman, David H. (2008) Explaining Increases in Higher Education Costs, The Journal of Higher Education, Vol. 79, No. 3 Baum, Sandy, and Payea, Kathleen (2003) Trends in Student Aid, College Board Publications. New York. Baumol, William J. and Sue Anne Batey Blackman (1989) “How to Think about Rising College Costs” Planning for Higher Education. 23, Summer, 1-7. Bowen, Howard R. (1980) The Cost of Higher Education: How Much Do Colleges and Universities Spend Per Student and How Much Should They Spend? San Francisco: Jossey-Bass. pp. 10-24 Cheslock, J. & Gianneschi (2008) Replacing State Appropriations with Alternative Revenue Sources: The Case of Voluntary Support. The Journal of Higher Education 79 (2). March/April 2008. Dill, D.D. (1997) Higher education markets and public policy, Higher Education Policy, 10, 167 – 185. Dimaggio, Paul J. and Walter W. Powell (1983). The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizations. Doyle, William R., Delaney, Jennifer A., and Blake Alan Naughton (2003) Public Institutional Aid and State Policy: Compensation or Compliance? Association for the Study of Higher Education Ehrenberg, Ronald G. (2002). Tuition Rising: Why College Costs So Much. Harvard University Press. Ehrenberg, Ronald G. (2002). Reaching for the Brass Ring: The U.S. News & World Report Rankings and Competition. The Review of Higher Education. Winter 2002 Vol. 26, No. 2. pp 145-162. Garvin, David A. (1980). The Economics of University Behavior, Academic Press. New York, NY. Pp 17-39. Gerald, Danette and Haycock, Kati (2007). Engines of Inequality: Diminishing Equity in the Nation’s Premier Public Universities. The National Education Trust. p. 7. 153 Getz, M., & Siegfried, J. J. (1991). Cost and productivity in American colleges and universities. In C. Clotfelter, R. Ehrenberg, M. Getz, & J. J. Siegfried (Eds.), Economic challenges in higher education (261–392). Chicago: University of Chicago Press. Greene, K.V. (1994) The public choice of non-resident college tuition levels, Public Choice, 78, 167 – 185. Griswold, Carolyn P. and Ginger M. Marine. (1996) “Political Influences on State Policy: Higher Tuition, Higher Aid, and the Real World” Review of Higher Education. 19(4). Summer. Pp. 361-389. Hearn J. C., & Griswold, C. P. (1994) State-level centralization and policy innovation in U.S. postsecondary education. Educational Evaluation and Policy Analysis, 16(2), 161 – 190. Hearn, Griswold, & Marine (1996). Region, Resources, and Reason: A Contextual Analysis of State Tuition and Student Aid Policies, Research in Higher Education, Vol. 37, No. 3 1996. pp 241-278. Hearn, James C. (1998) “The Growing Loan Orientation in Federal Financial Aid Policy” in ASHE Reader on Finance in Higher Education, 2nd Edition. J. Yaeger, G. Nelson, E. Potter, J. Weidman, and T. Zullo, eds. Boston: Pearson Custom Publishing. Hearn, James C. & David Longanecker (1985) “Enrollment Affects of Alternative Postsecondary Pricing Policies” Journal of Higher Education. 56(5). pp. 485-508. Heller, Donald. E. (1997) “Student Price Response in Higher Education: An Update to Leslie and Brinkman” Journal of Higher Education 68(6), Nov./Dec. pp. 624-659. Heller, Donald E. (2002). “The Policy Shift in State Financial Aid Programs” in Higher Education: Handbook of Theory and Research. Vol. 17. John C. Smart, ed. New York:Agathon Press. Pp 221-261. Hossler, D., Lund, J., Ramin, J., Westfall,, S. & Irish, S. (1997) State funding for higher education; The Sysyphean task, Journal of Higher Education, 68, 160 – 190. Hossler, D. & Anderson, D.K.(2004) The Enrollment Management Process. In Challenging and supporting the first-year student: A Handbook for improving the first year of college., edited by M.L. Upcraft, J.N. Gardner and B.O. Barefoot. San Francisco: Jossey-Bass. 154 Johnstone, D. Bruce. (1993). “The ‘High Tuition –High Aid’ Model of Public Higher Education Finance: The Case Against” paper from a presentation at the 1993 Annual Meeting of the National Association of Systems Heads. Johnstone, D. Bruce. (2004) “The Economics and Politics of Cost Sharing in Higher Education: Comparative Perspectives” Economics of Education Review. 23, pp.403-410. Leslie, L.L., & Brinkman, P.T. (1987). Student price response in higher education: The student demand studies. Journal of Higher Education, 58(2);, 181 – 204. Leslie, L.L. & Ramey G. (1986) State appropriation and enrollment: does enrollment growth still pay? Journal of Higher Education, 57, 1 – 19. Levy, Stanley R. (1995). “Sources of Current and Future Funding. New Directions for Student Services. No. 70. Summer. pp. 39-50. MacPherson, Michael S., and Morton O. Shapiro (1998) The Student Aid Game: Meeting Need and Rewarding Talent in American Higher Education. Princeton, NJ. Princeton University Press. McKeon, Howard P. (2003) “Controlling the Price of College” Chronicle of Higher Education. July 11. p. B20. McKeown, Mary P. (1982) “State Policies on Tuition and Fees for Public Higher Education” Journal of Education Finance, 8. (Summer 1982): 1-19. Rhoades, G. (1998) Managed Professionals: Unionized Faculty and Restructuring Academic Labor. Albany: State University of New York Press. Selingo, Jeffrey (2003) “The Disappearing State in Public Higher Education” Chronicle of Higher Education. February 28. Slaughter, Sheila, and Larry Leslie (1997) Academic Capitalism: Politics, Policies, and the Entrepreneurial University. Baltimore: Johns Hopkins University Press. Slaughter, Sheila, and Gary Rhoades (2004) Academic Capitalism and the New Economy. Baltimore: Johns Hopkins University Press. Tolbert, Pamela S. (1985) “Institutional Environments and Resource Dependence: Sources of Administrative Structure in Institutions of Higher Education.” Administrative Science Quarterly, 30. 155 Trow, Martin A. (1984) “The Analysis of Status,” in Burton R. Clark (Ed.), Perspectives on Higher Education: Eight Disciplinary and Comparative Views. Los Angeles: The University of California Press. US College Board, Trends in Higher Education Series, 2007. Weertz, David J., and Ronca, Justin M (2006) Examining Differences in State Support for Higher Education: A Comparative Study of State Appropriations for Research I Universities, The Journal of Higher Education, 77(6), 935 – 967. Wei, C.C., Li, X., and Berkner, L. (2004). A Decade of Undergraduate Student Aid: 1989-90 to 1999-2000 (NCES 2004-158). U.S. Department of Education, National Center for Education Statistics. Washington, D.C: U.S. Government Printing Office. Wellman, Jane V., Desrochers, Donna M. and Colleen M. Lenihan (2008). The Growing Imbalance: Recent trends in U.S. postsecondary education finance. Delta Project on Postsecondary Education Costs, Productivity and Accountability. Washington, D.C. pp 14-17. Wilkinson, Rupert (2004) Aiding Students, Buying Students: Financial Aid in America. Vanderbilt University Press. Winston, Gordon C. (1999). “Subsidies, Hierarchies, and Peers” Journal of Economic Perspectives. 13(1). Pp. 13-36. Wolanin, Thomas R. (1998) “Pell Grants: A 25-Year History” in Memory, Reason, Imagination: A Quarter Century of Pell Grants. L. Gladieux, ed. Washington, D.C.: College Board. pp. 13-31. Zemsky, Robert (1994). Faculty Discretionary Time: Departments and the Academic Ratchet. Journal of Higher Education 65, 1 – 22.
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project