|Title:||Exploring Washington's College Bound Scholarship Program: Sign-Up Decisions and the Relationship between Program Eligibility and Performance, Incarceration, College Going, and Persistence|
|Principal Investigator:||Goldhaber, Dan||Awardee:||American Institutes for Research (AIR)|
|Program:||Postsecondary and Adult Education [Program Details]|
|Award Period:||3 years (7/16/2014Ė7/15/2017)||Award Amount:||$691,345|
Previous Award Number:R305A140380
Co-Principal Investigator: Mark Long
Purpose: This project will address the income gap in college admission and completion by exploring Washington Stateís College Bound Scholarship Program, a program designed to improve college attainment for low-income Washington students. The Washington College Bound Scholarship Program (WSP) is a pledge-based scholarship program designed to help low-income students get on a path toward college and provide them resources to make college affordable. The WSP—and similar promise scholarship programs—differs from traditional need-based financial aid by eliminating the uncertainty associated with receiving the aid, and by making modest demands upon the student. For students who meet the demands—attaining a minimum 2.0 GPA during high school, not committing a felony, and completing a financial aid application during senior year—Washington State guarantees students an amount of financial aid that will fully cover their tuition, fees, and textbook costs when attending a Washington college. The goal of this research project is to determine if WSP encourages college enrollment and persistence. Researchers will use project findings to inform Washington policymakers about the participants in the intervention (WSP), and for whom it appears to be making a difference.
Project Activities: Researchers will collect administrative data for about 400,000 students in Washington State, as well as interview data from 60 principals and guidance counselors in the state. During the first half of the project, the research team will assess student and school characteristics associated with sign-up for the WSP program by eligible low-income students. During the second half of the project, the team members will assess whether WSP is associated with increased rates of college enrollment and persistence among low-income students.
Products: The products of this project include findings that will provide preliminary evidence of the promise of WSP for encouraging college enrollment and persistence among low-income Washington students. Peer-reviewed publications will also be produced.
Setting: This study includes all middle and high schools in the state of Washington.
Sample: Study participants include 396,290 middle school students in five cohorts from Fall 2005 to Fall 2009, with 144,622 of these eligible for WSP. The qualitative sub-sample will include a total of approximately 60 middle school principals and guidance counselors at schools attended by participating students.
Intervention: The Washington College Bound Scholarship Program (WSP) is a pledge-based scholarship program designed to help low-income students get on a path toward college and provide them resources to make college affordable. Students are eligible to sign the WSP pledge during 7th or 8th grade (or 9th grade for the first cohort during 2008–09) if any of the following applied: the student was eligible for free or reduced-price lunch (FRPL), the studentís family received Temporary Assistance for Needy Families (TANF), the student was a foster youth, or the studentís family income was below 185 percent of the poverty line ($39,220 in 2008). By signing the pledge during middle school, students commit to achieving at least a C average during high school, not committing a felony, and submitting the Free Application for Federal Student Aid during senior year of high school. If an eligible student signs the pledge, meets the obligations of the pledge, and his or her familyís income remains below a state-established threshold by her senior year, the student receives a scholarship that covers tuition and fees (plus a small textbook allowance) not covered by other state financial aid awards. The student can use the scholarship to attend any public or private Washington state higher education institution.
Research Design and Methods: The research team will use a mixed-method design that combines findings from semi-structured phone interviews with school principals and guidance counselors and findings from a multi-level regression model of WSP sign-up constructed from administrative data. The sampling strategy for interviews will guarantee coverage of schools with low and high sign-up and college-going rates. Interviews will be coded and analyzed using qualitative software, according to a thematic rubric aligned to the projectís research questions and interview protocol. A multi-level logistic regression model will be used to assess how individual and school characteristics predict the likelihood of students signing the pledge in middle school.
The analysis of college enrollment and persistence will employ a quasi-experimental differences-in-differences (DiD) design that compares college-going outcomes for students in pre- and post-WSP cohorts. By comparing differences in college-going for WSP-eligible and WSP-ineligible students across cohorts, the analysis will assess the likely effect of WSP on key college outcomes (e.g. enrollment and persistence). Additional specifications of the DiD model will assess possible moderators of the relation between WSP and college-going outcomes. To assess the possibility for positive peer influence, one model will assess whether WSP-outcome associations are greater in high schools that have high college-going and sign-up rates. To assess whether there is something about the sign-up process that encourages later success, a second model will assess whether WSP-outcome associations are greater for students who attended a middle school with a high sign-up rate.
Control Condition: The comparison group includes 161,703 students in two pre-program cohorts (2005-06 and 2006-07). These students did not receive program benefits because the Washington State legislature had not yet funded the program.
Key Measures: High school performance measures will include cumulative GPA, the difficulty level (according to Washington state guidelines) of a studentís course taking, on-time graduation, whether or not she received a diploma or GED, and whether or not she was ever incarcerated. Postsecondary attainment measures will include on-time (or delayed) enrollment, the sector and selectivity of the institution in which a student enrolls, and how many semesters she stays enrolled. Measures of gender, race/ethnicity, and region will facilitate the subgroup analysis.
Data Analytic Strategy: Researchers will use a mixed-method strategy to assess factors associated with students signing the pledge—to do well in high school, stay out of crime, and complete a Free Application for Student Aid during their senior year of high school. The data analysis will compare themes from qualitative interviews (with school principals and guidance counselors) to trends in the quantitative data indicating student and school characteristics associated with signing the pledge. The project will use a quantitative differences-in-differences model to assess relations between WSP and a host of college-going outcomes.
ERIC Citations:†Find available citations in ERIC for this award here or here.
Goldhaber, D., Long, M., Person, A., Rooklyn, J., & Gratz, T. (2019). Sign me up: The factors predicting students' enrollment in an early commitment scholarship program. AERA Open. doi.org/10.1177%2F2332858419857703. Full text
Goldhaber, D., Long, M. C., Gratz, T., & Rooklyn, J. (2020). Pledging to do "good": An early commitment pledge program, college scholarships, and high school outcomes in Washington State. Educational Evaluation and Policy Analysis, 42(1), 110–133. Full text
Long, M. C., Goldhaber, D., & Gratz, T. (2021). Washington's College Bound Scholarship Program and its effect on college entry, persistence, and completion. Education Finance and Policy, 16(4), 690–715.