Manajemen | Fakultas Ekonomi Universitas Maritim Raja Ali Haji 578.full

Financial Aid and Students’ College

Decisions

Evidence from the District of Columbia

Tuition Assistance Grant Program

Katharine G. Abraham

Melissa A. Clark

a b s t r a c t

The District of Columbia Tuition Assistance Grant Program (DCTAG), insti-tuted in 1999, allows D.C. residents to attend public colleges and universities throughout the country at considerably lower in-state tuition rates. We use the sharp decline in the price of public colleges and universities faced by D.C. residents to estimate the effects of price on college application and enrollment decisions. We find that DCTAG increased the likelihood that students applied to eligible institutions and markedly increased college enrollment rates among recent high school graduates. Enrollments increased primarily at less selective colleges and universities, with no decrease at more selective schools.

I. Introduction

How college costs influence students’ decisions about whether and where to attend is a question of considerable importance to policymakers, university Katharine Abraham is Professor of Survey Methodology and Adjunct Professor of Economics at the University of Maryland. Melissa Clark is an Economist at Mathematica Policy Research. The authors thank Orley Ashenfelter, Alan Krueger, David Linsenmeier, Cecilia Rouse, two anonymous referees, and seminar participants at the University of Notre Dame, the Federal Reserve Bank of New York, and the labor lunch at Princeton University for helpful comments and suggestions. Ellen Sawtell of The College Board and Diane Dickerson-Hayes of the University of the District of Columbia provided assistance with data. The SAT data used in this paper are proprietary and were obtained from the College Board under a license with Princeton University. Researchers wishing to obtain these data may apply for a license with the College Board, under the guidelines provided on their website: http://www.collegeboard.com/prod_ downloads/research/RDGuideforReleaseData.pdf. The authors are willing to provide guidance on using these data to anyone who may choose to pursue them. The other data used in this article can be obtained beginning January 2007 through December 2010 from Melissa Clark at Mathematica Policy Research, P.O. Box 2393, Princeton, NJ 08543 or mclark@mathematica-mpr.com.

[Submitted August 2005; accepted January 2006]

ISSN 022-166X E-ISSN 1548-8004 © 2006 by the Board of Regents of the University of Wisconsin System T H E J O U R NA L O F H U M A N R E S O U R C E S ● X L I ● 3

administrators, and education researchers alike. It is difficult, however, to determine the sensitivity of students’ application and enrollment decisions to the prices they face at various institutions. The amount of financial aid available to students typically is correlated with individual characteristics that may have an independent influence on their college decisions. Similarly, the price charged for attending particular colleges and universities is correlated with other institutional characteristics that may make them more or less attractive to potential enrollees. Drawing valid conclusions about the effect of college costs on students’ decisions requires an exogenous source of vari-ation in prices. In this paper, we utilize one such source of varivari-ation, the introduction of the District of Columbia Tuition Assistance Grant (DCTAG) program, to examine the effect of a sharp decline in public college and university prices faced by D.C. res-idents on their college application and enrollment decisions.

The DCTAG program, established in 1999, provides a substantial subsidy for D.C. residents to attend public colleges and universities outside the District. Specifically, the program allows qualifying D.C. residents to attend public colleges and universi-ties in other states at the considerably lower “in-state” tuition rates of these institu-tions. This benefit is subject to an annual cap of $10,000 and a lifetime cap of $50,000 per student. District of Columbia residents attending private institutions in the Washington, D.C. metropolitan area or designated historically black private colleges and universities are eligible for a smaller subsidy of $2,500 per year.

In this paper, we examine the effect of the DCTAG program on the college appli-cation and enrollment decisions of D.C. residents. We ask whether there are dis-cernible effects of the DCTAG program on the likelihood that a D.C. resident who graduates from high school will apply to college or actually enroll as a college fresh-man, and also whether the program has affected the type of colleges that D.C. high school graduates apply to or attend. We use these results to draw conclusions regard-ing the broader question of the effects of financial aid and the price of college on stu-dents’ college application and enrollment decisions.

Our work contributes to a growing literature on the effects of college price on stu-dents’ enrollment decisions. Much of the early research in this area failed to address the potential endogeneity of college prices faced by particular groups of students, but a handful of recent studies have exploited plausibly exogenous changes in the price of college faced by specific groups of students to study the effects of college price on college attendance.1_{These studies have provided important evidence of the effects of}

college price on enrollment decisions, but since most have exploited sources of vari-ation in college price that affect a specific subset of students, findings may not be eas-ily generalized to the broader U.S. population. For instance, Dynarski’s (2000) study of Georgia’s HOPE scholarship program and Kane’s (2003) study of the Cal Grant financial aid program exploit variation in college price due to merit-based financial aid, and findings are most appropriately generalized to higher performing students. Dynarski (2003) examines the 1983 elimination of the Social Security student bene-fit program, which had provided generous stipends to children of deceased parents aged 18 to 22 who were enrolled as full-time college students. Linsenmeier, Rosen,

1. See Leslie and Brinkman (1988) for a review of the early literature and Dynarski (2002) for an excellent overview of more recent studies.

and Rouse (2002) and Van der Klaauw (2002) each examine the effects of changes in financial aid policies at a particular university to see how these changes affect enroll-ments at that institution.

In another study of the DCTAG program, Kane (2004) examines the program’s overall effects on college enrollments, and also uses information about where students who fill out the Free Application for Federal Student Aid (FAFSA) chose to have that form sent to learn about DCTAG’s effects on college applications. A limitation of this approach is that many students do not complete the FAFSA.2_{Therefore, although}

interesting in its own right, analysis of FAFSA data is likely to paint an incomplete— and potentially distorted—picture of how the DCTAG program has affected college applications among all students eligible for the program.

Our study contributes to the existing literature in three main ways. First, we are able to examine the effects of college price on students from a broad socioeconomic spec-trum. Unlike most of the financial aid programs evaluated in previous work, DCTAG is neither need-based nor merit-based nor targeted at a specific population (such as children of deceased fathers), but is available to any D.C. resident meeting basic eli-gibility criteria, a group that is socio-economically diverse. Second, our study is one of the first that we know of to examine the effects of college price on college appli-cations as well as college attendance. In contrast to the FAFSA data analyzed by Kane (2004), we believe the SAT data we analyze provide a more comprehensive pic-ture of DCTAG’s effects on college applications. Moreover, we are able to examine how the program affects application decisions among a variety of subgroups of inter-est, including black versus white students, public versus private high school gradu-ates, and low versus high performing students. Third, our study examines how college price affects the selectivityof the schools chosen by students. The policy implications of an increase in enrollments at less selective public universities due to DCTAG, for example, would be quite different if the increase were due to a reduction in enroll-ments at top-tier private universities than if it were due to an increase in enrollenroll-ments among students who otherwise would have attended two-year colleges or not gone on to college at all. Our analysis of DCTAG’s effects on the selectivity of colleges applied to and attended provides valuable insight into the mechanisms through which college price affects college applications and enrollments.

We find that students are price-sensitive in their college application and enrollment decisions. By lowering the price of eligible public institutions, DCTAG increased the probability that students applied to, and enrolled at, these institutions. There is no evi-dence that DCTAG led students who would have otherwise attended more selective institutions to attend less selective schools eligible for the grant. Instead, the large

2. According to American Council on Education (2004), 50 percent of undergraduates enrolled for credit at eligible two-year or four-year institutions during the 1999–2000 academic year failed to complete a FAFSA. Full-time students are more likely than part-time students to fill out the form, but a full third of full-time undergraduates at public four-year colleges and more than 45 percent of those at public two-year colleges did not complete a FAFSA. Even among dependent students with family incomes under $20,000 per year, more than 20 percent did not complete a FAFSA, and noncompletion rates rise sharply with family income. In the District of Columbia in 1999, among families with children younger than 18, median family income was $33,757, and 91 percent of families had incomes greater than $20,000 (U.S. Census Bureau, 2005), sug-gesting that we can expect a significant share of D.C. high school graduates who go on to college not to have completed the FAFSA.

The Journal of Human Resources 580

growth in enrollments at eligible public institutions appears to have been fueled by increasing attendance rates among students who would not otherwise have gone to college. Overall, we find DCTAG increased the percentage of D.C. residents of high school graduation age who enrolled as college freshmen by roughly 8.9 percentage points or 3.6 percentage points for every $1,000 of aid.

II. Overview of the DCTAG Program

High school graduates in most states have access to a publicly funded higher education system that includes a network of colleges and universities. The only publicly funded institution of higher education in the District of Columbia is the University of the District of Columbia, an open-admissions institution that in many respects functions more like a community college than a typical university. Early in 1999, legislation to allow District high school graduates to attend public universities nationwide at in-state tuition rates was introduced in the U.S. House of Representatives. A more restrictive competing proposal, which would have limited program coverage to public institutions in Maryland and Virginia, soon appeared in the Senate. Those favoring tuition subsidy legislation argued that it would help to stem the outflow of middle class families, especially those with school age children, from the District and create invaluable opportunities for inner city youth. Given wide bipartisan support for the idea of tuition subsidies for D.C. high school graduates, passage of legislation in some form was widely anticipated.

The U.S. Congress passed the District of Columbia College Access Act of 1999 into law on November 12, 1999, and the first grants awarded under the program were for the 2000–2001 academic year. Under the legislation, any recent D.C. high school graduate who attends an eligible public institution may receive a benefit equal to the difference between that school’s in-state tuition and out-of-state tuition rates, up to $10,000 per year and $50,000 over the student’s lifetime. The law originally desig-nated all public institutions located in Maryland and Virginia as eligible institutions, and during the fall of 1999, when graduating high school seniors would have been applying to colleges for the fall of 2000, students most likely would have assumed that the DCTAG program would benefit only those attending Maryland or Virginia schools. But the law also allowed the Mayor of the District of Columbia to broaden the program’s scope if he or she determined that eligible students were experiencing difficulty in gaining admission to public colleges in Maryland and Virginia. In May 2000, the mayor took this step and formally extended the program to include all pub-lic institutions of higher education nationwide.

In addition to the subsidies provided for students attending public colleges and uni-versities, the law also provided smaller tuition subsidies for students attending private nonprofit colleges or universities located in the District of Columbia metropolitan area or private historically black colleges and universities located in Maryland and Virginia, and was later expanded to include all historically black colleges and univer-sities nationwide, effective with the 2002–2003 school year. Private school grants are subject to a maximum of $2,500 per year and $12,500 over a student’s lifetime.

To be eligible for benefits under the program, a student must have been a resident of the District of Columbia for at least 12 consecutive months prior to beginning his or

her freshman year of college; have graduated from high school or received a GED on or after January 1, 1998; have begun an undergraduate course of study within three years of high school graduation or receipt of the GED (extended for periods of mili-tary or Peace Corps service); be enrolled at least half-time in a course of study leading to an accredited degree; and not already hold a bachelor’s degree. In 2002 the eligibil-ity requirements were liberalized to allow older high school graduates to benefit from the program. The DCTAG program is not means tested; indeed, those who apply are not required to provide any information about their family’s income or assets.

Benefits may be applied only to tuition and fees, and benefit grants are to be made directly to the eligible institution, which must agree to comply with the program’s administrative requirements. In the DCTAG program’s first year, 1,900 individuals used the grant to attend 152 participating institutions (U.S. General Accounting Office 2002), and the number of students receiving awards had doubled by the program’s fourth year (Williams 2004). The number of students taking advantage of the program is large rel-ative both to the number of District residents of college age and to the number of District residents who graduate from high school: The Census Bureau estimates that, as of July 1, 2000, there were 5,765 17-year-olds resident in the District. In that same year, 2,695 stu-dents graduated from the District’s public high schools. (Comparable data on the number of District residents graduating from private high schools are not available.)

III. Data and Estimation Strategy

Previous research on the effects of variation in college costs typically has focused on how costs affect enrollments. Enrolling in college, however, is the end result of a sequence of steps. Students must identify a set of colleges they might like to attend; apply to some or all of those schools; be accepted for admission at one or more schools; and finally choose a particular school at which to enroll. Because we have access both to data on applications (as proxied by sending SAT scores to a school) and to data on enrollments, we are able to explore more fully than has previous research the effects of changes in costs on students’ decisions about college attendance.

A. SAT Dataset

To examine the college application decisions of individual students affected by the DCTAG program, we use data from the College Board on high school students who took the SAT. These data cover the 1994 through 2001 cohorts, where a cohort is defined as the group of students who on a normal schedule would graduate from high school in the indicated year. The database includes all black and Hispanic test takers, all test takers residing in Washington, D.C., and a 25 percent random sample of non-black, non-Hispanic test takers in other states. Most of the variables included in the database were derived from information supplied by the test takers on the Student Descriptive Questionnaire. Variables that we use in our analysis include colleges to which test scores were sent (four-digit code, up to 20 recorded in 1994–98 and up to 30 recorded in 1999–2001), state of residence, high school attended, type of high school attended (public, private, or other), race/ethnicity, highest level of education attained by mother and by father, and SAT math and verbal scores.

The Journal of Human Resources 582

We merge the data on each college or university to which students sent their scores to information on various characteristics of this institution provided by the College Board—whether it is a two- or four-year institution, whether it is public or private, and whether it is foreign or domestic. We also merge the data to measures of the selec-tivity of each institution. We define an institution as “top-tier” if it is ranked among the top-50 national universities or top tier of national liberal arts colleges according to the U.S. News and World Report 1998 College Rankings (U.S. News and World Report 1997).3

We treat a student’s decision to have SAT scores sent to a particular college or uni-versity as a proxy for the decision to apply to that institution. We expect that most four-year-college-bound high school seniors in D.C. and the other cities in our analy-sis will take the SAT and will send their SAT scores to every four-year college or uni-versity to which they apply. Sending SAT scores to a particular college is not a perfect proxy for actually applying to that college—some students may send their scores to colleges to which they decide ultimately not to apply or they may apply to colleges to which they do not send SAT scores (for instance, if the college does not require SAT scores for admission). As documented by Clark (2003), the vast majority of college-bound high school students in the District of Columbia and the other states repre-sented in our analysis take the SAT as opposed to the ACT. We are not aware of general information about the probability that a student who sends his or her SAT scores to a school will complete an application to the school, but Card and Krueger (2005) report that the number of students of a given race/ethnicity who send their SAT scores to particular public universities in the state of California is a good proxy for the number of students of a given race/ethnicity who actually apply to those same cam-puses. As most community colleges and other two-year colleges do not require SAT scores for admission, we limit our analysis of the SAT data to students applying to four-year colleges and universities.4

We adopt a difference-in-differences approach to study District students’ decisions about where to send their SAT scores, using students from large cities in nearby states (Baltimore, Maryland; Newark, New Jersey; Norfolk, Virginia; and Philadelphia, Pennsylvania) as a comparison group. To select these cities, we first identified nearby states with SAT-taking rates similar to those of students in the District of Columbia. The District has a population of 572,059 that is 60 percent African-American, and for our comparison group we identified cities in these states with populations of 200,000 or more that were at least 40 percent African-American, based on 2000 Census data

3. As an alternative measure, we define an institution as being in the top-tier if it is located in the United States; at least 500 students sent SAT scores to the institution in the year 2000; and the average SAT score of these applicants, as computed from the SAT database, exceeded the 75th percentile of scores in that year (1,160). As expected, there is considerable overlap between the lists of schools defined as top-tier institu-tions according to the two measures, and the two measures have a correlation of 0.74. Results are not qualitatively sensitive to which measure of selectivity we use, and we present only those using the selectiv-ity measure based on the U.S. News and World Report rankings.

4. An important consequence is that we are not able to examine DCTAG’s effects on students’ applications to two-year colleges. If the program induced students who otherwise would not have applied to college at all to apply to an eligible two-year college, that would not be reflected in this analysis. It would, however, be captured in our analysis of students’ enrollment decisions. We also note that the University of the District of Columbia does not require applicants to submit SAT scores.

as reported in U.S. Census Bureau (2001).5_{None of the states in which the}

compari-son cities are located instituted a college assistance program on anything like the scale of DCTAG over the period of interest. As the SAT dataset has information on stu-dents’ state of residence and high school (which can be matched to school address information), we identify SAT-takers by city based on their state of residence and their attendance at a high school located in the relevant city. The D.C. figures include D.C. residents who attend suburban high schools located in Maryland and Virginia.

Throughout our analysis, we exclude students without a valid SAT score and students who did not send their score to at least one four-year college or university. For consis-tency across years, we also exclude students who did not report an ethnicity (these stu-dents were missing from the 1994–98 samples provided by the College Board). Because we use high school address to match students to cities, students not reporting a valid high school code or students reporting a high school code that could not be matched to an address are excluded as well. Our results are not sensitive to any of these exclusions. Appendix Table A1 reports the means of selected characteristics for the SAT sample, both overall and then separately for D.C. versus the comparison cities.

We begin our analysis by examining trends in the number of students who send their SAT scores to at least one college or university, comparing the trend in the District of Columbia to that for the comparison cities. If the DCTAG program has had a positive effect on college applications, we would expect the share of D.C. students who send their SAT scores to at least one college to have grown relative to the share in the comparison cities.

Among those sending SAT scores to at least one school, we use the following model to identify the effect of the DCTAG program on wherestudents choose to send their scores:

(1) Yijt o Xijt t t ( DC .

t j j j t t t ijt

1 1995 2001

1995 2001

#

=c + c+ a d+ n z+ a i+y

= =

)

/

/

/

In this model, Yijtis the outcome variable of interest for student iin city jgraduating

from high school in year t; Xijtis a vector of individual-level characteristics; αt

repre-sents a set of year dummies, described below; µjrepresents a set of city dummy

vari-ables; DCis a dummy variable equal to one if the city is D.C., zero otherwise, and υijt

is a stochastic error term.

Outcomes examined using this model include whether student igraduating from high school in year t sent SAT scores to any DCTAG-eligible public Maryland or Virginia institution, whether the student sent scores to any DCTAG-eligible public institution anywhere in the United States, and various measures of the selectivity of the schools to which students sent scores. The Xijtvector includes individual-level

characteristics that are likely to be correlated with college application decisions, including race, parental education, and type of high school (public, private, or other) that the student attends.

The αtyear dummy variables are defined in the following way: α1995equals one if the year in which the student is expected to graduate from high school is 1995 or later

5. We also repeated the analysis using all test-takers from Virginia, Maryland, Connecticut, Delaware, New Jersey, North Carolina, and Pennsylvania as a comparison group. The results of this exercise are nearly iden-tical to those we report.

The Journal of Human Resources 584

and zero otherwise; α1996equals one if the student is expected to graduate in 1996 or

later and zero otherwise; and so forth, for each year from 1995 through 2001. Defined in this way, the coefficients on the year dummies represent the average change in the dependent variable (conditional on the covariates Xijt) in the comparison cities from

year t-1 to year t. The coefficient on the interaction between each αtand the D.C.

dummy variable represents the change in the dependent variable from year t-1 to year

tin D.C. relative to the comparison cities. For example, in a model of applications to DCTAG-eligible colleges, if DCTAG had a positive impact on the likelihood that a student applied to an eligible college, we would expect to see this reflected in posi-tive, significant coefficients on α2000* DCand/or α2001* DC(that is, in positive, sig-nificant values of θ2000and θ2001). These two coefficients represent the change in the

probability that a D.C. student applied to an eligible college between 1999 and 2000 and between 2000 and 2001, over and above any change in the probabilities for stu-dents in the comparison cities. The sum of the two coefficients (θ2000plus θ2001) rep-resents the total change in the probability of a D.C. high school graduate applying to an eligible college between 1999 and 2001.

B. IPEDS Dataset

To examine college enrollments of students from D.C. and surrounding states, we use data from the Integrated Postsecondary Education Data System (IPEDS), a dataset maintained by the National Center for Education Statistics of the U.S. Department of Education. The IPEDS contains data on institutions that grant baccalaureate or higher degrees, institutions that award two-year degrees, and less-than-two-year institutions. The IPEDS database consists of institution-level data collected through a set of annual surveys that together provide information on a range of items. In our analysis, we use institution-level IPEDS data for two- and four-year institutions from 1996, 1998, 2000, and 2002 on fall enrollments of freshmen by state of residence and in- and out-of-state tuition rates. We augment the IPEDS data with measures of college selectiv-ity; as described above, institutions are defined as “top-tier” based on their U.S. News and World Report college and university rankings for 1998.6

Like our analysis of the SAT data, our analysis of the IPEDS data employs a dif-ference-in-differences approach in which we examine changes in enrollment patterns among recent D.C. high school graduates following the introduction of the DCTAG program relative to the changes observed for a suitably-defined comparison group. For our SAT analysis, the comparison group consists of high school graduates residing in large, heavily African-American cities in nearby states. Because the IPEDS database identifies the number of college freshmen enrolled at reporting institutions by state but not by city of residence, our comparison group for the IPEDS analysis consists of recent high school graduates from the nearby states of Maryland, Virginia, Connecticut, Delaware, New Jersey, North Carolina, and Pennsylvania. Appendix Table A2 contains descriptive information regarding the schools contained in the IPEDS database.

6. Analyses of the IPEDS data using the alternative measure of selectivity based on average SAT scores described above produce qualitatively similar results and therefore are not reported.

Unfortunately, since the IPEDS does not contain information by state of residence on either the number of upperclassmen or the number of students who graduate, we are not able to examine DCTAG’s effects on retention or graduation rates, but only on the likelihood that students enroll as college freshmen. It would be of interest to know how the program affects students’ decisions to remain in college and their success in completing college, and it may be possible to address these questions in future research using different data sources.

IV. The Effects of DCTAG on College Application and

Enrollment Decisions

As already indicated, making the transition from high school to enrollment at a college or university is a multi-stage process during which a student faces multiple decision points. By changing students’ perceptions of affordability, the DCTAG program could encourage more students to consider college attendance and/or attendance at an eligible public institution, thereby increasing the number of applicants. The program also could affect enrollments by increasing the share of applicants who actually enroll as college freshmen. Our analysis examines the effects of the DCTAG program on outcomes at two decision points—the point at which students decide whether to apply to one or more colleges and the point at which they decide whether to enroll at a particular college.

A. DCTAG’s Effect on Applications to

Four-Year Colleges

One important objective of the DCTAG program is to encourage D.C. residents who would not otherwise have done so to apply to four-year colleges or universities. The total numberof D.C. residents sending SAT scores to a four-year college clearly increased following the introduction of the DCTAG program, from 1,394 in 1999 to 1,540 in 2001 (see Column 1 in the top panel of Table 1). As a fraction of the esti-mated population age 17 (Column 2), the share of District residents sending SAT scores to four-year colleges rose from 0.234 in 1999 to 0.281 in 2001 (Column 3), reversing what appears to have been a generally downward trend. We would also like to know whether the share of D.C. high school graduates who sent scores to four-year institutions rose following the introduction of DCTAG, but cannot calculate those figures because we do not know the number of District residents who graduate from private high schools each year.7_{We can, however, compare the number of D.C.} pub-lichigh school students sending SAT scores to four-year colleges with the number of The Journal of Human Resources

586

7. Data on high school graduation are collected from schools rather than students, and reporting by state is based on school location rather than student residence. In 1999, according to figures from the National Center for Education Statistics’ Common Core of Data and Private School Survey, 2,675 students graduated from the D.C. public high schools and 1,231 students graduated from private high schools located in the District. Many of the students who attend private high schools located in the District, however, are residents of Maryland or Virginia, and conversely, some D.C. residents attend private high schools located outside of the District’s boundaries in the Maryland and Virginia suburbs. While data on both residence and school

D.C. public high school graduates. Although the demographics of the public school population differ in some important ways from those of the full population of D.C. high school graduates, this group of students is of interest in its own right.8_Overall,

in each year from 1994 through 2001, roughly two-thirds of SAT-takers living in D.C. attended public high schools. The number of D.C. public high school graduates tak-ing the SAT increased considerably after the implementation of DCTAG, from 957 in 1999 to 1,084 in 2001 (Column 4). Although the number of students graduating from the District’s public high schools also increased over this period (Column 5), the share of District public high school graduates who took the SAT and had their scores sent to at least one four-year college increased by about 10 percent in 2000 and remained high in 2001 (Column 6), rising from 35.8 percent in 1999 to 38.6 percent in 2001.

In contrast to the experience in the District of Columbia, the number of SAT-takers actually fell in our comparison cities between 1999 and 2001, both overall (Column 1 in the bottom panel of Table 1) and among public high school students (Column 4). The share of public high school graduates in the comparison cities who took the SAT declined by about 6 percent over the same time period (Column 6), from 40.2 percent in 1999 to 37.7 percent in 2001.9_{These figures suggest that the DCTAG program encouraged}

stu-dents who might not otherwise have done so to apply to a four-year college.

The DCTAG program might be expected to have had an even larger effect on the

typesof four-year colleges to which D.C. residents apply, as it makes public colleges and universities located in other jurisdictions considerably more affordable.10_{In 1999,}

47.8 percent of D.C. SAT-takers sent their scores to at least one public four-year Maryland or Virginia school, and 73.5 percent sent their scores to at least one public four-year institution somewhere in the country (not shown in table). By 2001, these percentages had increased substantially, with 63.3 percent of D.C. SAT-takers send-ing their scores to at least one public Maryland or Virginia school, and 87.4 percent sending their scores to at least one public institution somewhere in the country.

location are not available for the full population of D.C. students, tabulations from the SAT database can pro-vide a general sense of high school enrollment patterns. Only 72 percent of SAT-takers who live in D.C. and attend a private high school attend a school located in the District, with 22 percent attending a Maryland high school and 5 percent a Virginia high school. Of the SAT-takers enrolled in private high schools located in the District, just 40 percent are District residents, 42 percent are Maryland residents, and 18 percent are Virginia residents. In contrast, 98 percent of SAT-takers who live in D.C. and attend a public high school attend a school located in the District, and 97 percent of those attending a D.C. public high school are District residents. 8. According to the National Center for Education Statistics, only 3 percent of those who graduated from the D.C. public high schools in 1999 were white, non-Hispanic, a share that is far below the share of white, non-Hispanics in the D.C. population of high school age. Observing both the classmates of the first author’s children, who attend D.C. public schools, and the classmates of their friends who attend various private schools, it is also apparent that the public school population is considerably less affluent on average than the private school population. Throughout this analysis, we categorize public high school graduates separately from charter school graduates.

9. The number of high school graduates is based on district-level reports contained in the National Center for Education Statistics’ Common Core of Data. City and school district boundaries coincide for each of the cities used in the analysis.

10. As already noted, the DCTAG program also offers a $2,500 subsidy to students who attend selected pri-vate colleges and universities. Most of these colleges and universities—particularly those that are included because they are located in the D.C. metropolitan area, rather than because they are historically black insti-tutions—are quite expensive, and the DCTAG subsidy is unlikely to make them affordable for someone who would have been deterred by their high tuitions in the first place.

The Journal of Human Resources 588

Table 1

Number of SAT Takers as Compared with the Number of 17-Year Olds and the Number of Public High School Graduates

Public High School SAT

All SAT Takers versus Takers versus

17-Year-Olds in Population Public High School Graduates

SAT 17-Year- SAT

Takers Olds Ratio Takers Graduates Ratio

(1) (2) (3) (4) (5) (6)

District of Columbia

1994 1,481 5,499 0.269 992 3,207 0.309

1995 1,520 5,415 0.281 1,060 2,974 0.356

1996 1,448 5,604 0.258 979 2,696 0.363

1997 1,451 5,765 0.252 990 2,853 0.347

1998 1,481 5,724 0.259 1,011 2,777 0.364

1999 1,394 5,956 0.234 957 2,675 0.358

2000 1,543 5,765 0.268 1,075 2,695 0.399

2001 1,540 5,484 0.281 1,084 2,808 0.386

Comparison cities

1994 9,974 — — 6,086 — —

1995 9,993 — — 6,263 — —

1996 10,350 — — 6,493 — —

1997 10,437 — — 6,561 — —

1998 10,805 — — 6,833 15,702 0.435

1999 10,348 — — 6,583 16,365 0.402

2000 9,831 — — 6,145 16,546 0.371

2001 9,917 — — 6,292 16,695 0.377

Source: SAT Score Database, Census Bureau, and National Center for Education Statistics Common Core of Data, various years.

Note: The SAT database contains information on SAT-takers sending scores to four-year colleges, and includes all black and Hispanic test takers; all test takers residing in Washington, D.C.; and a 25 percent ran-dom sample of nonblack, non-Hispanic test takers in other states. Number of SAT takers is an actual num-ber for D.C. and a weighted estimate for the comparison cities (Baltimore, Newark, Norfolk, and Philadelphia). Census Bureau population estimates are for July 1. Estimates for 1994–99 were adjusted to accord with overall growth in population measured by the 2000 Census. Population by year of age is not available for the comparison cities. Number of public high school graduates is based on district-level data from the Common Core of Data as reported by public school schools in the relevant cities. Each city con-tains only one regular public school district. Data on high school graduates from comparison cities are not available for 1994–97.

It is possible, of course, that other factors might have contributed to the observed increase in applications to out-of-state public institutions. Table 2 presents results from the difference-in-differences model specified in the preceding section and designed to account for this possibility by using SAT-takers from Baltimore, Newark, Norfolk, and Philadelphia as a comparison group. The year dummy variable coeffi-cients in these models capture the effects of any general trend influence on the propen-sity to apply to out-of-state public institutions. These models examine the propenpropen-sity of SAT-takers to send their scores to out-of-state institutions in Maryland or Virginia (Columns 1 and 2) or in any state (Columns 3 and 4).11_{The year dummies capture any}

general underlying trend in applications to qualifying schools; the coefficients on the interactions between the year dummies and the D.C. dummy represent the effects of anything that is D.C.-specific—most notably the DCTAG program—on District stu-dents’ application patterns. The models are fit first without any demographic controls and then with the addition of controls for a variety of demographic covariates that might be correlated with students’ college application decisions, including the type of high school a student attends (public, private, or “other,” where “other” includes char-ter schools and home schools, and represents less than 1 percent of the sample), the student’s race/ethnicity, the level of education obtained by the student’s parents, and the student’s combined math and verbal SAT score.

No consistent trend in the likelihood a student will send SAT scores to an out-of-state public institution is apparent in the simple year dummy coefficients. The coeffi-cients on the DC* year interactions, however, tell a clear story. Both in the models without demographic controls and in the models to which those controls have been added, the coefficient estimates for 2000 and for 2001 imply significant increases in the likelihood that a D.C. resident who took the SAT sent those scores either to a DCTAG-eligible institution in Maryland or Virginia or to a DCTAG-eligible institu-tion anywhere in the country.12

An important concern for education policymakers is how tuition subsidy and finan-cial aid policies will affect students from different socioeconomic backgrounds. In Table 3 we present results from our difference-in-differences model fit separately for various subgroups. For ease of interpretation, we present only the coefficients on the

DC* year interactions, which show the change in college application behavior of D.C. residents relative to their peers in the comparison cities and conditional on the demo-graphic characteristics described above.

We begin by examining the program’s effects on public versus private high school graduates.13_{For students in the District of Columbia, whether a student attends a public}

11. In the first pair of models, the dependent variable is equal to one for Maryland residents only if they applied to a DCTAG-eligible institution in Virginia, and for Virginia residents only if they applied to a DCTAG-eligible institution in Maryland. In the second pair of models, the dependent variable is equal to one only if the student applied to a DCTAG-eligible institution outside of his or her own state. We obtain results that are both qualitatively and quantitatively similar when Maryland and Virginia residents are excluded from the calculations. Results also are similar if the outcome variable is an indicator for whether the student applied to a public school in a neighboring state.

12. We also estimated a simpler difference-in-differences model that includes a single dummy variable for post-DCTAGin place of the year dummies, and a DC* post-DCTAGinteraction variable in place of the DC* yearinteractions, and obtained similar results.

13. Charter school students are not included in this analysis of public versus private high schools, but results are virtually identical if charter school students are included with public school students. Although the

Table 2

Determinants of Whether Students’ SAT Scores Sent to at Least One Out-of-State Public 4-Year College or University; Difference-in-Differences Model

Maryland or Virginia College College in Any State

(1) (2) (3) (4)

Constant 0.523 0.285 0.777 0.369

(0.013) (0.015) (0.011) (0.017)

1995 or later −0.019 −0.019 0.004 0.004

(0.007) (0.006) (0.010) (0.010)

1996 or later 0.020 0.020 0.012 0.011

(0.007) (0.007) (0.010) (0.010)

1997 or later −0.008 −0.009 −0.016 −0.018

(0.007) (0.007) (0.010) (0.010)

1998 or later −0.005 −0.003 0.010 0.014

(0.007) (0.006) (0.010) (0.010)

1999 or later −0.008 −0.010 −0.025 −0.027

(0.007) (0.007) (0.010) (0.010)

2000 or later −0.001 −0.002 0.012 0.010

(0.007) (0.007) (0.010) (0.010)

2001 or later 0.008 0.005 0.012 0.007

(0.007) (0.007) (0.010) (0.010)

DC*_{1995 or later −0.022 −0.020 −0.007 −0.005}

(0.019) (0.019) (0.018) (0.018)

DC*_{1996 or later −0.014 −0.016 −0.029 −0.031}

(0.020) (0.019) (0.019) (0.019)

DC*_{1997 or later 0.007 0.011 0.012 0.018}

(0.020) (0.020) (0.019) (0.019)

DC*_{1998 or later 0.041 0.039 0.014 0.013}

(0.020) (0.020) (0.019) (0.019)

DC*_{1999 or later −0.037 −0.036 −0.017 −0.016}

(0.020) (0.020) (0.019) (0.019)

DC*_{2000 or later 0.092 0.097 0.060 0.068}

(0.020) (0.020) (0.019) (0.019)

DC*_{2001 or later 0.056 0.057 0.055 0.057}

(0.019) (0.019) (0.017) (0.017)

State dummies yes yes yes yes

Demographic covariates no yes no yes

Observations 62,415 62,415 62,415 62,415

R-squared 0.13 0.15 0.09 0.14

DC*2000-or-later plus 0.148 0.155 0.115 0.126

DC*2001-or-later (0.019) (0.019) (0.018) (0.018)

Source: SAT Score Database.

Note: Sample includes SAT-takers graduating from high school in 1994 through 2001, sending scores to four-year colleges, and residing in Washington, D.C.; Baltimore, Md.; Newark, N.J.; Norfolk, Va.; or Philadelphia, Pa. Sample includes all black and Hispanic test takers; all test takers residing in Washington, D.C.; and a 25 percent random sample of nonblack, non-Hispanic test takers in other states. All models weighted OLS regressions. Heteroskedasticity-consistent standard errors in parentheses.

or a private high school is highly correlated with family socioeconomic status. Many D.C. residents who graduate from private high schools expect to attend prestigious private colleges but might also consider selected public universities. Interestingly, these students (or their parents) do appear to be price sensitive, at least with respect to the set of schools they will consider. Over the two-year period from 1999 to 2001, the share of college-bound D.C. private high school SAT-takers who sent their test scores to at least one Maryland or Virginia public college rose 19.5 percentage points (the sum of the DC*2000 or laterand DC*2001 or latercoefficients), controlling for other factors that might affect college applications. The increase for college-bound D.C. public high school SAT-takers was somewhat smaller, just 12.7 percentage points; however, the difference between the two subgroups was not statistically sig-nificant (p-value of difference = 0.118). The gap with regard to the share sending at least one application to any public university nationwide was larger and statistically significant, a 21.0 percentage point increase for the private high school graduates as compared with an 8.8 percentage point increase for the public high school graduates (p-value of difference = 0.003).

A similar pattern emerges for white takers as compared with black SAT-takers. The coefficients shown in Columns 5 and 6 imply that, holding other factors constant, the share of white SAT-takers applying to any Maryland or Virginia public college rose by 20.2 percentage points between 1999 and 2001, as compared with 14.2 percentage points for black SAT-takers (p-value of difference = 0.234). And the coefficients in Columns 7 and 8 imply that, again holding other factors constant, the share of white SAT-takers applying to any public college nationwide rose by 23.2 per-centage points, as compared with 10.3 perper-centage points for blacks (p-value of dif-ference = 0.013).

The program also appears to have a differential effect on students according to the educational attainment of their parents. Holding other factors constant, among SAT-takers with a parent who attended college, the share sending their scores to at least one Maryland or Virginia public college rose by 17.6 percentage points between 1999 and 2001, as compared with 8.2 percentage points for those whose parents completed no education beyond high school (p-value of difference = 0.022). The comparable fig-ures for applications to any public college nationwide are similar, showing an increase of 14.9 percentage points for those who had at least one parent who attended college, as compared with an increase of 4.5 percentage points for those whose parents never went beyond high school (p-value of difference = 0.004).

Differences between D.C. SAT-takers with above-average and below-average SAT scores are somewhat less marked. In the college-bound group with combined math and verbal scores above 1000, the share who sent at least one application to a public Maryland or Virginia school rose 17.1 percentage points between 1999 and 2001, as compared with an increase of 13.8 percentage points among those with combined scores of 1000 or less, but the difference is not statistically significant. The share

District of Columbia has among the highest charter school enrollment of any state, during the time period of our analysis, charter school students constituted less than 1 percent of D.C. SAT-takers. Charter school enrollment has grown over time, but even among the 2001 cohort, only 4 percent of D.C. SAT-takers were enrolled in a charter school.

applying to at least one public college nationwide rose 16.6 percentage points among those with combined scores above 1000, as compared with 9.6 percentage points among those with lower combined scores (p-value of difference = 0.077).

Overall, the results from Table 3 suggest that DCTAG had a greater impact on the probability of applying to at least one eligible four-year colleges among students from more advantaged backgrounds. One reason for this may be that these students were less likely to have considered public colleges and universities prior to DCTAG, mean-ing that there was more scope for the program to have altered their application behav-ior. In 1998, only 65.4 percent of D.C. SAT-takers from private high schools sent their scores to one or more public colleges or universities, compared with 81.2 percent of D.C. SAT-takers from public high schools, and there were similar differences in the pre-DCTAG application behavior of white as compared with black takers, SAT-takers with college-educated parents as compared with those whose parents had not attended college, and SAT-takers with above-average as compared with below-average test scores. In interpreting our results, it also is important to keep in mind that the SAT data allow us to look only at applications to four-year colleges and universi-ties. It is possible that DCTAG had an equal or greater impact on the likelihood that students from less advantaged backgrounds applied to any eligible two- or four-year institution. Although we cannot examine applications to two-year colleges with the SAT data, we will examine enrollments at these institutions in our analysis of the IPEDS data.

A related question is how DCTAG affected the selectivity of colleges to which students apply. Does the program, for example, encourage students who otherwise would have applied only to top-tier four-year private colleges instead to apply to eli-gible lower-tier public four-year colleges? We examine this question in various ways in Table 4. In 1999, before DCTAG took effect, 55.1 percent of D.C. SAT-takers sent their scores to at least one top-tier school, as identified based on the U.S. News and World Report rankings. On average, these students sent their scores to 1.71 top-tier schools and 4.00 lower-tier schools. As shown in the first column of Table 4, DCTAG appears to have had no effect on the probability that SAT-takers sent their scores to at least one top-tier college, which remained fairly constant over this period. DCTAG also appears to have had no effect on the average number of top-tier colleges to which D.C. residents applied (Column 2); the increase in the average number of total applications between 2000 and 2001 appears to have been driven almost entirely by a 0.262 increase in the average number of applications to lower-tier schools (Column 3). DCTAG slightly decreased the fraction of total scores students sent to top-tier colleges, although this change is not statistically sig-nificant (Column 4).

In sum, DCTAG appears to have increased the number of students applying to four-year colleges and universities; increased the likelihood that college-bound students applied to DCTAG-eligible four-year colleges and universities; and increased the number of applications to less selective colleges, while not significantly decreasing the number of applications to top-tier colleges. It also appears to have had a greater effect on the college application decisions of students from more advantaged back-grounds. Although DCTAG affected where students applied, however, it may not have affected whether or where they actually enrolled. To examine this question, we turn next to a discussion of DCTAG’s effects on college enrollment.

The Journal of Human Resources 592

B. DCTAG’s Effect on College Enrollments

One of the goals of the DCTAG program is to increase college enrollments of D.C. residents. All of the data we have examined thus far relate to the decision to applyto a (four-year) college or university, as captured by the decision to send SAT scores to a school. In this section we examine data on college enrollments from the IPEDS database. In particular, we focus on the number of D.C. residents who graduated from high school in the previous 12 months who are enrolled as first-time college freshmen in fall 1996, 1998, 2000 or 2002.14_{The IPEDS data provide an important complement}

to the SAT data, allowing us to examine where D.C. residents actually choose to enroll after the introduction of DCTAG, including enrollments at both two- and four-year colleges.15

In Table 5, we report enrollments for 1996, 1998, 2000 and 2002 for those schools that had the largest increases and the largest declines in enrollment of D.C. residents between 1998 and 2002, together with summary figures on enrollments at DCTAG-eligible and non-DCTAG-DCTAG-eligible schools. The school that appears to have benefited most from the introduction of the DCTAG program is Virginia State University, a his-torically black public institution at which the enrollment of D.C. freshmen jumped by 92 students between 1998 and 2000, to 108 students, and remained at that higher level in 2002. Altogether 16 schools added 10 or more D.C. freshmen to their enrollment counts between 1998 and 2002; 13 of these 16 schools were DCTAG-eligible institu-tions. In addition to Virginia State University, other historically black public colleges with notable D.C. enrollment gains included North Carolina Agricultural and Technical University, Norfolk State University (in Virginia), the University of Maryland Eastern Shore, Coppin State College (in Maryland) and North Carolina Central University. Fewer schools experienced D.C. freshman enrollment declines of more than ten students, but three of the six that did, including the two schools with the largest decline in enrollment of District residents, are private institutions located in the District of Columbia.16

Summary figures on enrollments of D.C. freshmen are shown at the bottom of Table 5. When the DCTAG program was first introduced, students planning to start college in the fall of 2000 were told that it would benefit students enrolled in Maryland and Virginia public institutions. Between fall 1998 and fall 2000, freshmen enrollments of D.C. residents at these schools nearly tripled, jumping from 162 stu-dents to 449 stustu-dents. This increase came partially at the expense of enrollments at

14. The IPEDS has information on both total freshmen enrollment and enrollment of freshmen who gradu-ated from high school in the past 12 months. We focus on the latter because, at least initially, only recent high school graduates were eligible for the DCTAG program.

15. Although in principle the IPEDS database contains information on enrollments for all schools in every year, reports for some schools are missing in some years. The tabulations shown in our tables are based on a balanced panel of schools that provided reports for 1996, 1998, 2000, and 2002. We obtain very similar results when we include data for schools that reported in some but not all years. The enrollment data for the University of the District of Columbia (UDC) contained in the IPEDS database are obviously inconsistent across years. Rather than use this information, we have substituted corrected figures obtained directly from UDC’s Office of University Statistics.

16. Students attending Howard University, George Washington University, or the other private D.C. schools listed earlier are eligible for a smaller DCTAG tuition break of $2,500, but as noted previously, this subsidy is small relative to the cost of attending these institutions.

The Journal of Human Resources

594

Table 3

Determinants of Whether Students’ SAT Scores Sent to at Least One Out-of-State Public 4-Year College or University; Difference-in-Differences Models, Selected Subgroups

Type of High School Attended Race

Maryland or College Maryland or College Virginia College in Any State Virginia College in Any State Public Private Public Private White Black White Black

(1) (2) (3) (4) (5) (6) (7) (8) DC*1995 or later −0.001 −0.069 0.006 −0.033 −0.018 −0.007 −0.045 0.018

(0.023) (0.033) (0.021) (0.035) (0.040) (0.023) (0.047) (0.019) DC*1996 or later −0.012 −0.005 −0.012 −0.049 −0.023 −0.009 −0.060 −0.021

(0.023) (0.033) (0.021) (0.036) (0.039) (0.023) (0.047) (0.020) DC*1997 or later 0.025 −0.026 0.027 −0.011 −0.001 0.007 0.050 0.010

(0.024) (0.033) (0.021) (0.037) (0.038) (0.024) (0.046) (0.020) DC*1998 or later 0.002 0.120 −0.009 0.062 0.079 0.023 0.009 0.007

(0.023) (0.034) (0.021) (0.036) (0.041) (0.024) (0.047) (0.020) DC*1999 or later 0.002 −0.125 0.013 −0.093 −0.081 −0.014 −0.063 −0.017

(0.024) (0.035) (0.021) (0.037) (0.044) (0.024) (0.050) (0.020) DC*2000 or later 0.091 0.098 0.046 0.110 0.053 0.088 0.069 0.059

(0.023) (0.035) (0.020) (0.038) (0.043) (0.023) (0.050) (0.020) DC*2001 or later 0.036 0.097 0.042 0.099 0.149 0.054 0.163 0.044

(0.022) (0.037) (0.018) (0.036) (0.045) (0.022) (0.047) (0.017) Year change dummies yes yes yes yes yes yes yes yes State dummies yes yes yes yes yes yes yes yes Demographic covariates yes yes yes yes yes yes yes yes Observations 46,972 15,178 46,972 15,178 10,458 44,245 10,458 44,245 R-squared 0.19 0.12 0.15 0.13 0.08 0.15 0.09 0.11 DC*2000-or-later plus 0.127 0.195 0.089 0.210 0.202 0.142 0.232 0.102

DC*2001-or-later (0.023) (0.037) (0.019) (0.037) (0.045) (0.023) (0.049) (0.019) Difference between subgroups −0.067 −0.122 0.060 0.130 P-value of difference [0.118] [0.003] [0.234] [0.013]

Abraham and Clark

595

Maryland or College Maryland or College Virginia College in Any State Virginia College in Any State

High High Greater Greater

school College school College 1,000 than 1,000 than or less or more or less or more or less 1,000 or less 1,000

(9) (10) (11) (12) (13) (14) (15) (16) DC*1995 or later −0.041 −0.007 0.028 −0.010 −0.025 −0.012 −0.002 −0.011

(0.037) (0.022) (0.031) (0.022) (0.023) (0.033) (0.020) (0.036) DC*1996 or later −0.021 −0.015 −0.008 −0.040 0.001 −0.027 −0.011 −0.055

(0.039) (0.022) (0.032) (0.022) (0.024) (0.033) (0.020) (0.036) DC*1997 or later 0.046 −0.008 −0.026 0.027 0.024 −0.018 0.014 0.030

(0.038) (0.023) (0.033) (0.023) (0.024) (0.032) (0.021) (0.036) DC*1998 or later 0.010 0.048 0.022 0.009 0.018 0.059 0.018 −0.019

(0.037) (0.023) (0.032) (0.023) (0.024) (0.033) (0.020) (0.036) DC*1999 or later 0.017 −0.055 0.017 −0.026 −0.038 −0.029 −0.027 0.017

(0.037) (0.023) (0.032) (0.023) (0.024) (0.034) (0.021) (0.037) DC*2000 or later 0.037 0.114 0.001 0.088 0.110 0.070 0.072 0.051

(0.035) (0.023) (0.031) (0.023) (0.023) (0.034) (0.020) (0.037) DC*2001 or later 0.044 0.062 0.044 0.061 0.028 0.101 0.024 0.115

(0.032) (0.023) (0.027) (0.021) (0.022) (0.035) (0.018) (0.035) Year change dummies yes yes yes yes yes yes yes yes State dummies yes yes yes yes yes yes yes yes Demographic covariates yes yes yes yes yes yes yes yes Observations 21,009 41,406 21,009 41,406 47,603 14,812 47,603 14,812 R-squared 0.25 0.12 0.18 0.11 0.22 0.07 0.18 0.08

The Journal of Human Resources

596

Table 3 (continued)

Highest Parental Educational Attainment SAT Verbal plus Math Score Maryland or College Maryland or College Virginia College in Any State Virginia College in Any State

High High Greater Greater

school College school College 1,000 than 1,000 than or less or more or less or more or less 1,000 or less 1,000

(9) (10) (11) (12) (13) (14) (15) (16) DC*2000-or-later plus 0.082 0.176 0.045 0.149 0.138 0.171 0.096 0.166

DC*2001-or-later (0.034) (0.023) (0.029) (0.022) (0.023) (0.035) (0.019) (0.036) Difference between subgroups –0.094 –0.104 −0.033 −0.072 P-value of difference [0.022] [0.004] [0.429] [0.077]

Source: SAT Score Database.

Note: Sample includes SAT-takers in specified subgroups graduating from high school in 1994 through 2001, sending scores to four-year colleges, and residing in Washington, D.C.; Baltimore, Md.; Newark, N.J.; Norfolk, Va.; or Philadelphia, Pa. Sample includes all black and Hispanic test takers; all test takers residing in Washington, D.C.; and a 25 percent random sample of nonblack, non-Hispanic test-takers in other states. All models weighted OLS regressions. Heteroskedasticity-consistent standard errors in parentheses.

other schools – the number of D.C. freshmen fell both at public colleges in other states and at private colleges – but the total number of recent D.C. high school graduates enrolled as college freshmen grew by 154 students. Between 2000 and 2002, there were large D.C. enrollment gains at DCTAG-eligible schools in other states; enroll-ments at DCTAG-eligible schools in Maryland and Virginia dropped off slightly and enrollments at non-DCTAG-eligible institutions rebounded to about their 1998 level.

Table 4

Determinants of Number and Share of Scores Sent to Top-Tier Schools

Sent Scores _{Number of Scores Sent to} Fraction of

to at Least Scores Sent

One Top- Top-tier Nontop-tier to Top-tier

tier School Schools Schools Schools

(1) (2) (3) (4)

DC*1995 or later −0.006 0.063 −0.097 0.004

(0.019) (0.085) (0.097) (0.010)

DC*1996 or later 0.005 0.147 0.047 0.014

(0.019) (0.087) (0.098) (0.010)

DC*1997 or later −0.000 0.006 −0.098 0.000

(0.019) (0.088) (0.099) (0.010)

DC*1998 or later −0.006 −0.136 0.167 −0.010

(0.019) (0.085) (0.099) (0.010)

DC*1999 or later 0.013 0.000 0.072 0.002

(0.019) (0.083) (0.105) (0.011)

DC*2000 or later −0.016 −0.057 0.043 −0.014

(0.019) (0.084) (0.104) (0.011)

DC*2001 or later 0.008 0.041 0.219 −0.005

(0.019) (0.084) (0.102) (0.010)

Year change dummies yes yes yes yes

State dummies yes yes yes yes

Demographic covariates yes yes yes yes

Observations 62,415 62,415 62,415 62,415

R-squared 0.18 0.32 0.02 0.31

DC*2000-or-later plus −0.008 −0.015 0.262 −0.019

DC*2001-or-later (0.019) (0.083) (0.105) (0.010)

Source: SAT Score Database.

Note: Sample includes SAT-takers graduating from high school in 1994 through 2001, sending scores to four-year colleges, and residing in Washington, D.C.; Baltimore, Md.; Newark, N.J.; Norfolk, Va.; or Philadelphia, Pa. Sample includes all black and Hispanic test takers; all test takers residing in Washington, D.C.; and a 25 percent random sample of nonblack, non-Hispanic test-takers in other states. All models weighted OLS regressions with heteroskedasticity-consistent standard errors. Top-tier schools are defined as those appearing in the top tier (top-50) of the U.S. News and World Report1998 “Best National Universities” list or the top tier (top-42) of the U.S. News and World Report 1998 “Best National Liberal Arts Colleges” list.

The Journal of Human Resources

598

Table 5

Fall Enrollment of D.C. Freshmen Completing High School in Last 12 Months, by School

Number of D.C. Freshmen Change in Number of D.C. Freshmen

2- or DCTAG State of 1996 1998 2000 2002 1996–98 98–2000 2000–02 98–2002 4-year eligible? school

School name (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

1 Virginia State University 19 16 108 108 −3 92 0 92 4 1 VA

2 University of the 287 260 245 309 −27 −15 64 49 4 0 DC

District of Columbia

3 Trinity College 55 26 31 65 −29 5 34 39 4 0 DC

4 Virginia Commonwealth 4 2 22 33 −2 20 11 31 4 1 VA

University

5 North Carolina A&T 12 9 13 39 −3 4 26 30 4 1 NC

6 Norfolk State University 22 12 60 39 −10 48 −21 27 4 1 VA

6 Montgomery College of 11 9 27 36 −2 18 9 27 2 1 MD

Rockville

8 University of Pittsburgh- 0 0 9 25 0 9 16 25 4 1 PA

Main Campus

9 University of Maryland- 25 18 29 38 −7 11 9 20 4 1 MD

Eastern Shore

10 Temple University 6 17 14 35 11 −3 21 18 4 1 PA

11 George Mason

University 12 3 16 20 −9 13 4 17 4 1 VA

11 University of 0 0 4 17 0 4 13 17 4 1 WI

Wisconsin-Madison

Abraham and Clark

599

Community College

14 North Carolina 9 3 7 14 −6 4 7 11 4 1 NC

Central University

14 Georgetown University 10 13 5 24 3 −8 19 11 4 0 DC

364 Morgan State

University 13 14 29 2 1 15 −27 −12 4 1 MD

365 Clark Atlanta

University 19 21 9 7 2 −12 −2 −14 4 0 GA

366 Strayer University- 34 31 22 7 −3 −9 −15 −24 4 0 DC

Washington Campus

367 Hampton University 10 34 21 8 24 −13 −13 −26 4 0 VA

368 Howard University 123 86 29 48 −37 −57 19 −38 4 0 DC

369 George Washington 86 60 23 17 −26 −37 −6 −43 4 0 DC

University

Total: MD and VA 172 162 449 410 −10 287 −39 248 DCTAG-eligible

schools

Total: Other DCTAG- 247 304 245 485 57 −59 240 181 eligible schools

Total: All DCTAG- 419 466 694 895 47 228 201 429

eligible schools

Total: Non-DCTAG 1,336 1,182 1,108 1,176 −154 −74 68 −6 eligible schools

Total: All schools 1,755 1,648 1,802 2,071 −107 154 269 423

Source: Integrated Post-Secondary Education Data System Database.

Note: Tabulations are based on the set of two- and four-year institutions in Maryland, Virginia, Connecticut, Delaware, New Jersey, North Carolina, and Pennsylvania for which the IPEDS database contains valid records for the years 1996, 1998, 2000, and 2002.

On net, between 1998 and 2002, enrollments of D.C. freshmen at DCTAG-eligible institutions grew by 429 students. Over the same period, overall enrollments of D.C. freshmen rose by 423 students, an increase of more than a quarter in the number of D.C. high school graduates going on to college the following year.

One potential concern about DCTAG is that recent D.C. graduates could be choos-ing less-selective public institutions in preference to the top-tier private institutions they otherwise would have attended. There is, we would note, some question about whether this would in fact be a bad thing. Results reported by Dale and Krueger (2002) suggest that, on average, controlling for an individual student’s observed and unobserved characteristics, the selectivity of the college the student chooses to attend has little effect on subsequent earnings, though they also find that college selectivity may be more important for students from low income families. In any event, the data we have examined provide no reason to think that DCTAG has significantly affected enrollments at top-tier private institutions. Table 6 reports enrollments of D.C. fresh-men at two-year colleges, lower-tier four-year colleges and top-tier four-year colleges, broken out separately for DCTAG-eligible and non-DCTAG-eligible institutions. As before, we use the U.S. News and World Report rankings of colleges and universities to determine whether a school is a top-tier institution. Most of the increase in enroll-ments at DCTAG-eligible institutions occurred at lower-tier four-year colleges, but there were also smaller increases at two-year and top-tier four-year DCTAG-eligible schools. Overall, the number of students enrolled at “better” schools (at four-year rather than two-year colleges, and at top-tier institutions within the four-year group), rose following DCTAG’s introduction.

To put the college enrollment increases observed among District high school grad-uates into perspective, we would like to compare the enrollment trends observed in the District with those observed in our comparison states. Whereas Census Bureau estimates show the high school age population of the District to have been stable or declining, that in the comparison states has been growing. In Table 7A, we compare the change in college enrollments as a share of the population of 17-year-olds in D.C. with that in our comparison states.17_{After falling between 1996 and 1998, the percent}

of D.C. 17-year-olds going on to college rose by 10.1 percentage points between 1998 and 2002, from 28.8 percent in 1998 to 38.9 percent in 2002; most of this increase was due to increases in enrollments at DCTAG-eligible schools. In contrast, in the comparison states, college enrollments as a fraction of the population age 17 increased by just 1.3 percentage points, from 42.1 percent in 1998 to 43.4 percent in 2002. The difference-in-differences estimate based on these figures implies that DCTAG was associated with an 8.9 percentage point increase in the ratio of college enrollments to population age 17.

Another way to look at the data would be to ask how the share of high school grad-uates who enroll in college changed in D.C. versus the comparison states following the introduction of DCTAG. As noted earlier, we do not know the number of D.C. res-idents graduating from private high schools, and therefore cannot simply calculate the

17. Using the number of 18- or 19-year-olds as the denominator for this share would be problematic, as many people of these ages move to the District from elsewhere and would not be eligible for DCTAG. In 2000, for example, there were 5,860 17-year-old District residents, about the same as the number of residents for each year of age from 12 through 16, but there were 8,911 18-year-olds and 11,903 19-year-olds. The Journal of Human Resources

Abraham and Clark

601

Fall Enrollment of D.C. Freshmen Completing High School in Last 12 Months, by Year and Type of Institution

Number of D.C. Freshmen Change in Number of D.C. Freshmen 1996 1998 2000 2002 1996–1998 1998–2000 2000–2002 1998–2002

(1) (2) (3) (4) (5) (6) (7) (8)

DCTAG-eligible

Two-year 42 36 76 76 −6 40 0 40

Four-year, nontop-tier 361 399 586 764 38 187 178 365

Four-year, top-tier 16 31 32 55 15 1 23 24

Total 419 466 694 895 47 228 201 429

Non-DCTAG-eligible

Two-year 7 6 17 2 −1 11 −15 −4

Four-year, nontop-tier 1,021 871 762 849 −150 −109 87 −22

Four-year, top-tier 308 305 329 325 −3 24 −4 20

Total 1,336 1,182 1,108 1,176 −154 −74 68 −6

All schools

Two-year 49 42 93 78 −7 51 −15 36

Four-year, nontop-tier 1,382 1,270 1,348 1,613 −112 78 265 343

Four-year, top-tier 324 336 361 380 12 25 19 44

Total 1,755 1,648 1,802 2,071 −107 154 269 423

Source: Integrated Post-Secondary Education Data System Database.

Note: Tabulations are based on the set of two- and four-year institutions in Maryland, Virginia, Connecticut, Delaware, New Jersey, North Carolina, and Pennsylvania for which the IPEDS database contains valid records for the years 1996, 1998, 2000, and 2002.

The Journal of Human Resources

602

Table 7a

Freshmen Enrollment as a Share of Population of 17-Year Olds, by Year and Type of Institution

Difference D.C. Comparison Comparison and Comparison D.C. Level D.C. Change State Level State Change State Changes

1996– 1998– 1996– 1998– 1996– 1998–

1996 1998 2002 1998 2002 1996 1998 2002 1998 2002 1998 2002

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)

DCTAG-eligible

Two-year 0.7 0.6 1.4 −0.1 0.8 0.2 0.3 0.2 0.1 0.0 −0.2 0.8

Four-year, nontop-tier 6.4 7.0 14.4 0.5 7.4 3.6 3.5 3.9 −0.1 0.3 0.6 7.1 Four-year, top-tier 0.3 0.5 1.0 0.3 0.5 0.3 0.3 0.3 0.0 0.0 0.2 0.5

Total 7.5 8.1 16.8 0.7 8.7 4.1 4.1 4.4 0.0 0.3 0.7 8.4

Non-DCTAG-eligible

Two-year 0.1 0.1 0.0 0.0 −0.1 8.3 9.7 9.9 1.4 0.2 −1.4 −0.2

Four-year, nontop-tier 18.2 15.2 16.0 −3.0 0.7 23.5 24.3 25.1 0.8 0.8 −3.8 −0.1 Four-year, top-tier 5.5 5.3 6.1 −0.2 0.8 3.8 3.9 3.9 0.1 0.0 −0.3 0.8

Total 23.8 20.6 22.1 −3.2 1.5 35.6 38.0 38.9 2.4 1.0 −5.6 0.5

All schools

Two-year 0.9 0.7 1.5 −0.1 0.7 8.5 10.0 10.1 1.5 0.1 −1.6 0.6

Four-year, nontop-tier 24.7 22.2 30.3 −2.5 8.1 27.1 27.9 29.0 0.7 1.1 −3.2 7.0 Four-year, top-tier 5.8 5.9 7.1 0.1 1.3 4.0 4.2 4.2 0.2 0.0 −0.1 1.3 Total 31.3 28.8 38.9 −2.5 10.1 39.7 42.1 43.4 2.4 1.3 −4.9 8.9

Source: Integrated Post-Secondary Education Data System Database and Census Bureau.

Abraham and Clark

603

Percentage Growth in Enrollment of College Freshmen Less Percentage Growth in Number of Public High School Graduates, by Year and Type of Institution

D.C. Excess Enrollment Comparison State Excess Difference D.C. and Growth Rate Enrollment Growth Rate Comparison State Rates 1996–1998 1998–2002 1996–1998 1998–2002 1996–1998 1998–2002

(1) (2) (3) (4) (5) (6)

DCTAG-eligible

Two-year −17.3 99.8 42.2 −22.9 −59.5 122.7

Four-year, nontop-tier 7.5 80.5 −3.2 5.4 10.7 75.0

Four-year, top-tier 90.7 62.9 9.2 4.7 81.6 58.2

Total 8.2 80.8 −0.3 3.5 8.5 77.3

Non-DCTAG-eligible

Two-year −17.3 −77.9 17.2 −2.7 −34.5 −75.3

Four-year, nontop-tier −17.9 −14.4 3.3 −1.0 −21.3 −13.4

Four-year, top-tier −2.3 −2.4 3.8 −5.3 −6.2 2.9

Total −14.5 −11.8 6.6 −1.9 −21.2 −9.9

All schools

Two-year −17.3 74.4 17.8 −3.2 −35.1 77.7

Four-year, nontop-tier −11.4 15.1 2.5 −0.2 −13.8 15.3

Four-year, top-tier 2.6 3.9 4.2 −4.6 −1.7 8.5

Total −9.1 14.4 5.9 −1.4 −15.0 15.8

Source: Integrated Post-Secondary Education Data System Database and Digest of Education Statistics, various years.

C. Estimates of the Effect of College Price on Enrollment and the Cost of Added Enrollment

For comparison with the literature, it is useful to calculate the effect on the college attendance rate of $1,000 of aid provided by the DCTAG program. To determine the responsiveness of D.C. students’ college enrollment decisions to the change in col-lege costs associated with the DCTAG program, we need first to estimate the value of the DCTAG subsidy to the typical participating student. We use information on in-state and out-of-state tuitions contained in the IPEDS database for this purpose. We approx-imate the average potential savings in college costs for District residents following the introduction of DCTAG by calculating the amount that someone enrolling at each school represented in the IPEDS database would have saved in 2000 and then aver-aging those amounts, weighted by the enrollment of D.C. freshmen at each school in 2000.20 _{For public institutions outside the District, DCTAG reduces costs by the} smaller of $10,000 and the difference between in-state and out-of-state tuition and for designated private schools, by $2,500; for all other private schools, costs are unchanged. The average savings in college costs for D.C. residents, calculated in the fashion just described, was $2,472. The estimates in Table 7A imply that DCTAG was associated with an 8.9 percentage point increase in the fraction of 17-year-olds choos-ing to enroll in college. These estimates of change in cost and change in enrollments imply roughly a 3.6 percentage point increase in the enrollment rate per $1,000 reduc-tion in the cost of college. This estimate, although rough, is consistent with the esti-mated effects in the literature for similarly transparent policy and program changes.

In their summary of the earlier literature, Leslie and Brinkman (1988) conclude that each $100 (in 1982-83 dollars) in the cost of college reduces college enrollments by 0.6 to 0.8 percentage points; these numbers imply that a $1,000 (in 2000 dollars) reduction in the cost of college should raise enrollment rates by about 3.5 to 4.5 per-centage points.21_{It is important to note, however, that the estimates from these earlier} studies may suffer from selection bias, and may not accurately reflect the true causal effect of price on enrollments.

More recent studies that exploit exogenous variation in college price report similar effects. The estimates reported by Dynarski (2000) imply that each $1,000 (in 2000 dollars) of assistance under the Georgia HOPE Scholarship program, introduced in

20. For consistency with the data on enrollments we have examined, this estimate of average potential tuition savings is based on data for schools for which the IPEDS database contains information for all four of the years 1996, 1998, 2000, and 2002. The question of how to calculate the change in college costs faced by D.C. residents when enrollment patterns have changed in response to the change in relative cost of attend-ing public versus private institutions is akin to the more standard question of how to calculate a cost of liv-ing index when the bundle of goods and services purchased by consumers is not fixed. Usliv-ing enrollment patterns prior to the introduction of DCTAG, when fewer students attended eligible public institutions, to cal-culate average cost savings is analogous to computing a Laspeyres price index and would understate the true reduction in college costs; using enrollment patterns observed after the shift in enrollments to eligible pub-lic institutions has been completed is analogous to computing a Paasche price index and would overstate the true reduction in college costs. Using the 2000 enrollment patterns, which reflect some but not all of the pro-gram’s effects on where D.C. students go to college, gives an answer intermediate between the two extremes. 21. For the purpose of these calculations, earlier year dollars were converted to 2000 dollars using the Consumer Price Index. If, for example, the CPI had doubled between a given year and 2000, a $500 cost change in the earlier year would be comparable to a $1,000 cost change in the year 2000.

1993, increased the college attendance rate at eligible schools in Georgia by 3.5 to 4.0 percentage points. Dynarski (2003) estimates that the elimination in the early 1980s of the $6,700 (in 2000 dollars) in Social Security survivors’ benefits available to chil-dren of deceased parents who were full-time college students reduced enrollment rates among that group by 20 to 25 percentage points, implying that each $1,000 in added cost reduced enrollments by roughly 3 to 4 percentage points. As noted earlier, both of these studies provide evidence on the effects of college price relevant to the population affected by the programs under study—students with a high school grade average of B or above in the case of Dynarski (2000) and students with deceased par-ents in the case of Dynarksi (2003). Our own estimates, based on the recent introduc-tion of the DCTAG program, reflect the effect of college price on enrollments of a broader spectrum of students, and may therefore be more readily generalizable to the population at large.

Given the structure of the DCTAG program and that the grant was used by many stu-dents who likely would have attended college even had the program not existed, the cost incurred per student induced to attend college who would not otherwise have done so clearly exceeds the average subsidy to participating students. Part of the reason for the program’s success seems likely to be its universal coverage and the fact that appli-cants are not required to expend much effort to establish their eligibility for the tuition assistance grant. Still, it seems worth thinking about the cost per extra enrollee under the program. Turning again to our estimates from Table 7A, an increase in college enrollment of 8.9 percentage points among an estimated population of 5,321 17-year-olds implies that, in 2002, DCTAG was responsible for 474 students attending college who would not otherwise have done so. The total tuition subsidy received by students who enrolled as freshmen under the program in that year was $6,753,782, calculated by summing the available subsidies for all D.C. freshmen attending eligible colleges in 2002. This means that the cost of the program per additional student enrolled, ignor-ing any administrative overhead costs, was about $14,250. Whether it is a good invest-ment to spend this amount of money to raise college enrollinvest-ments depends, among other things, on the returns to the additional education obtained, something we unfortunately cannot estimate with any degree of confidence.

V. Conclusions

The DCTAG program provides an unusual opportunity to study the effects of a large, exogenous change in tuition price on students’ decisions about applying to and then enrolling in college. This program reduced the costs of college for D.C. residents significantly. Following the program’s introduction, the share of D.C. high school graduates who apply to four-year colleges appears to have risen somewhat, and the share who enroll as college freshmen has risen markedly. Further, the DCTAG subsidy has encouraged D.C. high school graduates who are considering college to apply to eligible public colleges and universities, especially those located in Virginia and Maryland, and subsequently to attend these same schools. Our study of the DCTAG program provides important evidence that should be of interest to pol-icymakers considering tuition assistance programs and to university administrators considering their school’s financial aid policies.

Our analysis suggests that students are price-sensitive in their college application decisions. We find that, by lowering the relative price of public institutions, DCTAG significantly increased the probability that D.C. SAT-takers applied to at least one pub-lic four-year college or university. DCTAG also appears to have increased the number of applications to lower-tier four-year colleges while not significantly decreasing the number of applications to top-tier institutions, suggesting that students who would have applied to top-tier schools in the absence of DCTAG continue to do so under the program, but apply in addition to lower-tier schools eligible for the grant. Among those applying to at least one four-year college, DCTAG had a considerably larger effect on the probability of applying to at least one eligible school for SAT-takers from more advantaged backgrounds—for students attending a private high school versus those attending a public one, for white students versus black students, for students with a parent that attended college versus those whose parents did not, and (to a lesser degree) for students scoring higher on the SAT versus those scoring lower on the exam. It is important to keep in mind, however, that we are only able to examine appli-cations to four-year colleges and universities; any effect DCTAG may have had on the propensity to apply to eligible two-year colleges among less advantaged students is not reflected in this analysis.

We also find that price has a significant effect on whether and where students decide to enroll in college. The IPEDS data provide strong evidence that DCTAG affected the pattern of college enrollments among D.C. residents at both two- and four-year colleges and universities. Total enrollment of D.C. high school graduates as college freshmen increased by 25 percent between 1998 and 2002, the third year of the DCTAG program. Freshmen enrollments of D.C. residents at eligible Maryland and Virginia schools increased by more than 150 percent between 1998 and 2002, and nearly doubled at eligible schools throughout the country over the same period. There is no evidence that DCTAG led students who would have attended more selective institutions instead to attend less selective schools eligible for the grant; indeed, enrollments at top-tier four-year institutions (as identified by U.S. News and World Report) rose slightly over the 1998 to 2002 time period, rather than falling. The larger growth in enrollments at lower-tier institutions over the same time period appears to have been fueled instead by increasing attendance rates among students who would not otherwise have gone on to college.

Overall, DCTAG raised the percentage of D.C. residents of high-school-graduation age who enrolled as freshmen by roughly 8.9 percentage points, or 3.6 percentage points for every $1,000 of aid. This estimate, although roughly comparable to oth-ers in the literature, is less vulnerable to selection bias than some previous esti-mates, given that it is based on a plausibly exogenous source of variation in college price. Further, the DCTAG experience has the virtue of being relatively recent and, because program benefits are available to students from a diverse socioeconomic spectrum, more readily generalizable to the U.S. population at large than estimates from studies that examine effects of policy changes targeted at specific population subgroups.

Like other public policies and programs designed to alter the costs of college that have been found to have significant effects on college enrollments, the DCTAG program is simple to understand and highly visible to D.C. residents who might be eligible for its benefits. While the specific circumstances that led to the introduction of the DCTAG

Table A1

Descriptive Statistics for SAT Database (Percentages Unless Otherwise Indicated)

Overall D.C. Comparison Cities

Female 58.04 59.76 57.79

Race/ethnicity

White 37.99 17.73 40.93

Black 47.31 69.78 44.05

Hispanic 2.55 5.06 2.18

Other 12.15 7.42 12.84

Parental education

High school or less 33.28 26.29 34.30

At least one parent 66.72 73.71 65.70

some college or more Type of high school

Public 63.52 68.71 62.77

Private 35.97 30.31 36.80

Other 0.50 0.98 0.43

Average SAT verbal

plus math score 911.0 930.1 908.3

City of residence

Washington, D.C. 12.68 100.00 0.00

Baltimore, Md. 21.77 0.00 24.93

Newark, N.J. 7.57 0.00 8.67

Norfolk, Va. 6.26 0.00 7.17

Philadelphia, Pa. 51.72 0.00 59.23

Sample size 62,415 11,858 50,557

Source: SAT Score Database.

Note: Sample includes SAT-takers graduating from high school in 1994 through 2001, sending scores to four-year colleges, and residing in Washington, D.C.; Baltimore, Md.; Newark, N.J.; Norfolk, Va.; or Philadelphia, Pa. Sample includes all black and Hispanic test takers; all test takers residing in Washington, D.C.; and a 25 percent random sample of nonblack, non-Hispanic test takers in other states. Weighted estimates.

program and the details of the program’s design are unique to the District of Columbia, the fact that this program has been so successful in encouraging college attendance among students drawn from an inner city, largely minority population should help to inform the decisions of policymakers at both the national and the state level who would like to encourage increased college attendance among similar populations.

Table A2

Characteristics of Institutions in IPEDS Database, Weighted by Freshmen Enrollment from States and Years in Sample (Percentages Unless Otherwise Indicated)

Weighted by Freshmen Enrollment from

All States in Comparison

Sample D.C. States

Control

Public 68.7 49.1 68.8

Private 31.3 50.9 31.2

Level

Two-Year 22.9 3.6 23.0

Four-Year 77.1 96.4 77.0

Top-tier 13.8 20.5 13.8

Historically black 4.4 43.2 4.3

In-state tuition (dollars)

1996 6,800 9,158 6,792

1998 7,443 10,069 7,434

2000 7,370 10,065 7,361

2002 8,350 11,164 8,340

Out-of-state tuition (dollars)

1996 10,042 11,261 10,037

1998 10,864 12,315 10,859

2000 11,013 12,483 11,008

2002 12,568 13,925 12,563

Number of institutions in sample 1,973 566 1,969

with freshmen enrollment from specified state(s) in 1996, 1998, 2000, or 2002

Total freshmen enrollment from 252,626 1,819 250,807

specified state(s) (average over 1996, 1998, 2000, 2002)

Source: Integrated Post-Secondary Education Data System Database, various years, and U.S. News and World Report(1997).

Note: Tabulations are based on the set of two- and four-year institutions for which the IPEDS database con-tains valid records for the years 1996, 1998, 2000 and 2002, and are weighted by the total freshmen enroll-ment in 1996, 1998, 2000, and 2002 from the specified state(s) of residence. Comparison states are Maryland, Virginia, Connecticut, Delaware, New Jersey, North Carolina, and Pennsylvania. Top-tier schools are defined as those appearing in the top tier (top-50) of the U.S. News and World Report 1998 “Best National Universities” list or the top tier (top-42) of the U.S. News and World Report1998 “Best National Liberal Arts Colleges” list.

References

American Council on Education. 2004. Missed Opportunities: Students Who Do Not Apply for Financial Aid. Issue Brief, Washington, D.C., October.

Card, David, and Alan B. Krueger. 2005. “Would the Elimination of Affirmative Action Affect Highly-Qualified Minority Applicants? Evidence from California and Texas.” Industrial and Labor Relations Review58(3):416–34.

Clark, Melissa A. 2003. “Selection Bias in College Admissions Test Data.” Working Paper, Princeton University, Princeton, New Jersey.

Dale, Stacy Berg, and Alan B. Krueger. 2002. “Estimating the Payoff to Attending a More Selective College: An Application of Selection on Observables and Unobservables,” Quarterly Journal of Economics117(4):1491–527.

Dynarski, Susan. 2000. “Hope for Whom? Financial Aid for the Middle Class and Its Impact on College Attendance.” National Tax Journal53(3):629–62.

———. 2002. “The Behavioral and Distributional Implications of Aid for College.” American

Economic Review92(2):279–85.

———. 2003. “Does Aid Matter? Measuring the Effect of Student Aid on College

Attendance and Completion.” American Economic Review93(1):279–88.

Kane, Thomas J. 2003. “A Quasi-Experimental Estimate of the Impact of Financial Aid on College-Going.” NBER Working Paper #9703.

———. 2004. “Evaluating the Impact of the D.C. Tuition Assistance Grant Program.” NBER

Working Paper #10658.

Leslie, Larry, and Paul Brinkman. 1988. The Economic Value of Higher Education. New York: Macmillan.

Linsenmeier, David M., Harvey S. Rosen, and Cecilia Elena Rouse. 2002. “Financial Aid Packages and College Enrollment Decisions: An Econometric Case Study.” NBER Working Paper #9228.

National Center for Education Statistics. 1999. Indirect State-Level Estimation for the Private School Survey. Technical Report NCES 1999-351. Washington, D.C.

U.S. Census Bureau. 2001. County and City Data Book: 2000. Washington, D.C.: GPO.

———. 2005. “American FactFinder Detailed Tables.” Available at http://factfinder.census.gov

(accessed December 19, 2005).

U.S. General Accounting Office. 2002. D.C. Tuition Assistance Grants: Program May Increase College Choices, but a Few Program Procedures May Hinder Grant Receipt for Some Residents. Washington, D.C.: GPO.

U.S. News and World Report. 1997. “America’s Best Colleges 1998.” September, pp. 79–114. Van der Klaauw, Wilbert. 2002. “Estimating the Effect of Financial Aid Offers on College

Enrollment: A Regression-Discontinuity Approach.” International Economic Review 43(4):1249–87.

Williams, Anthony A. 2004. Testimony on the reauthorization of the D.C. College Access Act before the Committee on Government Reform, U.S. House of Representatives, March 25.