Co-Principal Investigator: Liang Zhang (Pennsylvania State University)

Purpose: In this study, the researchers will evaluate the effects of Florida's Bright Futures Scholarship program, a merit aid program, on: (1) college enrollment and degree production in the state and its different postsecondary institution types, and (2) students' college choice, persistence, and degree completion.

Project Activities: Secondary data analyses will be carried out on both macro- and micro-level data. Institution-level data from the Integrated Postsecondary Educational Data System (IPEDS) will be used to examine Bright Futures, college enrollment, and degree production in Florida overall by type of institution, gender, and race/ethnicity. Student-level data from the Florida Educational Data Warehouse (FEDW) will be used to examine Bright Futures and individual student college choice, persistence, and degree completion.

Products: Researchers will provide evidence on the efficacy of the Bright Futures Scholarship program and report this evidence in peer reviewed journals.

Structured Abstract

Setting: The study will take place in Florida.

Population: The data come from Florida students attending colleges in each academic year from 1995–96 through 2005–06. Data on each cohort will contain student education progression in the education system from high school, to college, and to graduation.

Intervention: Florida's Bright Futures Scholarship program, started in 1997, combined two existing programs in the state, Florida Academic Scholars Award (FAS) for students on academic tracks and Florida Gold Seal Vocational Scholars Award (GSV) for students on vocational tracks, and added Florida Medallion Scholars Award (FMS), also for students on academic tracks. The FAS awards cover 100 percent of a student's tuition fees and there is also some allowance for fees and college-related expenses. The program requires a 3.5 GPA on 15 college preparatory credits in high school and an SAT of 1270 or ACT score of 28 for initial qualification. A 3.0 GPA on all postsecondary work attempted is required for renewal. The FMS awards cover 75 percent of tuition and required fees while requiring 3.0 GPA on 15 college preparatory credits and SAT at 970 or ACT at 20 for initial qualification. A 2.75 GPA on all postsecondary work is required for renewal. The GSV awards are similar to FMS award but are for students in vocational tracks (Florida Department of Education, 2010). The current study focuses on students on academic tracks, so only FAS and FMS are relevant for this study.

Research Design and Methods: A difference-in-differences method will be used to examine the effects of merit aid programs on college enrollment and degree production in Florida and other states. The annual IPEDS enrollment survey data from every postsecondary institution will be aggregated up to the state level including the creation of such outcome variables as freshman enrollment and degrees conferred. Data from other sources will be used to identify other states with merit aid programs (and the dates of those programs) and to create control variables such as the number of recent high school graduates, state per capital personal income, and state unemployment rates. Statistical analysis of the data will compare differences in changes in the outcomes before and after the establishment of state merit aid programs with changes in states that did not institute merit aid programs.

To examine student level outcomes, both a regression discontinuity design and a pre- and post- policy differences design will be used with the FEDW data on students attending college from 1995–96 through 2005–06. The first will compare students who were just above the eligibility score (GPA for college students and GPA and SAT/ACT score for high school students) for the scholarship versus those just below it. Two types of pre- and post-policy difference designs will be used. One will examine differences in outcomes for Florida college students before and after Bright Futures was implemented. The second will be difference-in-differences designs by adding a comparison group (e.g., non-Florida students who attend Florida postsecondary schools but are not eligible for Bright Futures; Florida students not eligible for Bright Futures either before or after it was implemented).

Control Condition: For the macro-level analysis, states without a merit aid scholarship but that are similar in other ways to Florida will be the comparison group. For the micro-level analyses, students just below the scholarship eligibility criteria will be the comparison in the regression discontinuity design.

Key Measures: Measures include aggregated institutional data from IPEDS to examine the overall impact of Bright Futures program on college enrollment and degree production in Florida. Dependent variables will be created using FEDW student-level data on topics including college choice (selectivity of the institution that the student attended, according to Barron's), college persistence (whether students returned to college from year to year), and college graduation (whether students received a college degree). Financial aid information includes whether or not the student received a Bright Futures scholarship for each semester, the type and amount of Bright Future scholarship each semester, and the type and amount of other financial aid information, if available. Data on students' academic performance include high school transcript data, high school grades, high school class rank, SAT and/or ACT scores, college transcript data, college grades for each course, college GPAs at the end of each semester, and credit hours enrolled each semester. FEDW also provides information on family background information (e.g., mother's education, father's education, eligibility for free or reduced lunch) and student demographics (e.g., gender, race/ethnicity, age, residency status).

Data Analytic Strategy: The macro-level analysis will use feasible generalized least squares to address autocorrelation from the previous year. To address other differences among states that may impact the outcome variables, state-level covariates will be included along with both state and time fixed effects. In addition, different comparison groups (composed of states) will be used to check for robustness of results. Local linear regression around the cutoff point will be used with the regression discontinuity design. However, to address potential non-compliance, 2-stage least squares will also be estimated with an instrumental variable used to predict the probability of using the scholarship, and this probability will be used to estimate the effect of the treatment on the treated.

Project website: http://www.coe.fsu.edu/bfrp

Products and Publications

Book chapter

Hu, S., Trengove, M., and Zhang, L. (2012). Toward a Better Understanding of the Effects of State Merit Aid Programs. In J. Smart, and M. Paulsen (Eds.), Higher Education: Handbook of Theory and Research (pp. 291–334). New York: Springer.

Journal article, monograph, or newsletter

Zhang, L., Hu, S., and Sensenig, V. (2013). The Effect of Florida's Bright Futures Program on College Enrollment and Degree Production: An Aggregated-Level Analysis. Research in Higher Education, 54: 746–764.