Skip Navigation
Funding Opportunities | Search Funded Research Grants and Contracts

IES Grant

Title: Evaluation of a Random Assignment Intervention to Improve College Choice Among High Achieving, Low Income Students
Center: NCER Year: 2010
Principal Investigator: Hoxby, Caroline Awardee: Stanford University
Program: Postsecondary and Adult Education      [Program Details]
Award Period: 4 years Award Amount: $2,879,635
Type: Efficacy and Replication Award Number: R305A100120
Description:

Co-Principal Investigator: Sarah Turner (University of Virginia)

Purpose: Evidence suggests that low-income, high-achieving students are not as likely to attend selective colleges as their high-income, high-achieving peers. However, selective colleges have much greater resources per student than typical nonselective four-year colleges, including such resources as more faculty members per student, better trained faculty, and better campus facilities and services. This study evaluates the effects of four interventions designed to increase application to selective colleges by qualified low-income students through providing informational and financial supports.

Project activities: This project will identify high-achieving high school students in low-income families who attend high schools that traditionally do not send high-achieving students to selective colleges using data from the College Board, ACT, and U.S. Census. These students will be invited to join the study and high school seniors with consent will be randomly assigned to receive one of four interventions or to a control group. High school students are surveyed in the spring of their senior year to gather information on participation in the interventions, college applications, and college plans. In the spring of their college freshman year they are surveyed again to gather information on their first year college experience. Data from the College Board and ACT will be used to determine which colleges the students applied to while data from the National Student Clearinghouse will be used to determine which college a student enrolls in and their persistence in postsecondary education.

Products: The products of this project will be published articles in peer reviewed journals describing the interventions and their impact on college application and enrollment in selective colleges for high-achieving, low-income students.

Structured Abstract

Setting: The project is national in scope and targets high-achieving students in high schools least likely to send students to selective colleges and universities.

Population: The population is high-school students who are low income, high achieving, attending high schools that do not send their students on to selective colleges, and members of the prospective college freshman classes of 2010 and 2011. Approximately 20,000 seniors will be randomly selected from SAT and ACT test-takers in fall of 2009 and fall of 2010.

Intervention: The four interventions in the project were selected based on promising results from previous studies to provide different resources and supports to encourage application to selective colleges. One intervention gives students information on application strategies. A second intervention gives students information on the net-of-financial aid costs they can expect to pay. The third intervention gives students a set of no-paperwork fee waivers for application to state flagship universities and other selective colleges and universities. The fourth intervention provides information to parents on application strategies and financial aid, with particular attention to explaining how college choice may affect family finances and how to support their student's aspirations to attend a selective college.

Research Design and Methods: Upon receipt of agreement to participate in the study, students will be randomly assigned to control status or one of the interventions. Before implementing the randomization, students will be grouped by state then randomly assigned in order to allow estimation of the effect of moderating variables (such as the in-state cost of the public flagship university) that vary systemically by state.

Control Condition: The control condition consists of students whose parents returned the consent form but were not randomly selected to participate in the program.

Key Measures: Two sets of key measures will be used. The first set are used identify the students to be sampled: 1) high-achieving is defined as scoring at or above the 90th percentile on the SAT or ACT and self-reporting a grade point average of at least A-, 2) low-income is defined as being in the bottom quartile of family income which is estimated from U.S. Census data at the block level and IRS current income data at the zip code level, and 3) high schools with high rates of low-income students applying to selective colleges are identified using College Board and ACT data and their students are excluded from the study. The second set are the outcomes used to evaluate whether any of the four interventions increase the likelihood that students will: (a) apply to a range of colleges that include selective ones; (b) obtain a wider range of admissions and financial aid offers; (c) enroll at colleges that are more selective, have more plentiful resources, or provide more generous financial aid; and (d) complete the freshman year of college. Data for these are obtained from the National Student Clearinghouse, SAT and ACT testing services, and student surveys. Data Analytic Strategy: Regression analysis will be used to analyze continuous outcome variables (e.g., number of colleges applied to) while probit and linear probability models will be used with binary outcome variables (e.g., apply to any selective college). Conditional logistic regression will be used for multiple choice outcomes such as the college where a student chooses to enroll (the explanatory variables will be college characteristics and college-student match specific characteristics). Moderators will be examined through interactions between the treatment indicator variables and the moderator variables, e.g., student gender, state characteristics, high school and neighborhood characteristics.

Intervention: The four interventions in the project were selected based on promising results from previous studies to provide different resources and supports to encourage application to selective colleges. One intervention gives students information on application strategies. A second intervention gives students information on the net-of-financial aid costs (the amount of direct costs after financial aid is applied) they can expect to pay. The third intervention gives students a set of no-paperwork fee waivers for application to state flagship universities and other selective colleges and universities. The fourth intervention provides information to parents on application strategies and financial aid, with particular attention to explaining how college choice may affect family finances and how to support their studentís aspirations to attend a selective college.

Research Design and Methods: Students who agree to be in the study will be randomly assigned to control status or to one of the four interventions. Students will first be grouped by state then randomly assigned. Stratifying the randomization by state is to increase the likelihood of estimating the effect of moderating variables (such as the in-state cost of the public flagship university) that vary systemically by state.

Control Condition: Three control conditions will be used for: (1) students whose parents received and returned the informed consent request ("control respondents") and were randomly assigned to the control group; (2) students whose parents received but did not return the informed consent request ("control non-respondents"; and, (3) students excluded from the parental consent request ("control non-participants"). Including the latter two control groups will allow the project to check on the degree of selection into the random assignment design.

Key Measures: The project will evaluate whether any of the four interventions increase the likelihood that students will: (a) apply to a range of colleges that include selective ones; (b) obtain a wider range of admissions and financial aid offers; (c) enroll at colleges that are more selective, have more plentiful resources, or provide more generous financial aid; and (d) complete the freshman year of college. Data are obtained from the National Student Clearinghouse, SAT and ACT testing services, and student surveys.

Data Analytic Strategy: Utilizing the randomized design, regression techniques will be applied to predict college related academic outcomes that are coded in binary and multiple levels.

Publications

Journal article, monograph, or newsletter

Hoxby, C.M., and Turner, S. (2015). What High-Achieving Low-Income Students Know About College. The American Economic Review, 105(5), 514–517.

Working paper

Hoxby, C.M., and Avery, C. (2012). The Missing ";One-Offs";: The Hidden Supply of High-Achieving, Low Income Students (NBER 18586). Cambridge, MA: National Bureau of Economic Research Working Paper.


Back