Co-Principal Investigator: Meredith Phillips and Sara Reber (University of California, Los Angeles)

Purpose: The research team will test the efficacy of V-SOURCE (Virtual Student Outreach for College Enrollment), a 15-month program designed to increase college access among academically eligible, disadvantaged high school students by providing support for the college application process. The project will also test the efficacy of a less intensive variant of the program, Milestones, which focuses on getting students to complete key college-application milestones using automated reminders and incentives. Both programs are based on an earlier program known as SOURCE that used one-on-one mentoring to support students as they applied for college and was tested in a prior randomized control trial.

Project Activities: Researchers will implement an experiment to test the efficacy of the V-SOURCE and Milestones programs. Students will be randomly assigned to one of three groups. The first group is to receive the V-SOURCE program, the second group is to receive the Milestones program, and the third group will be the control group. Two cohorts of juniors in Southern California high schools serving large percentages of low-income and minority students will be recruited to take part in the study. Approximately 80 eligible students from each of 40 schools will be recruited in each cohort for a total of approximately 6,400 students. A baseline survey, given before randomization, will collect information on students' demographic and academic backgrounds, plans for the future, educational expectations and expectations for returns to education, and understanding of the college application process. Additional student information (e.g., standardized test scores, GPA, SAT-taking and scores, eligibility for attending a 4-year California public college or university) will be collected from district data. Outcome data on student enrollment in postsecondary education and semesters completed will be obtained from the National Student Clearinghouse. The results of the experiment will be analyzed to determine the overall impact of the two programs as well as impacts on subgroups including gender, parental education, parental language, and students' time preferences.

Products: Products include evidence of the efficacy of V-SOURCE and Milestones published in peer-reviewed journals.

Structured Abstract

Setting: The study will take place in several urban and outlying school districts in Southern California, including the Los Angeles Unified School District (LAUSD), that serve a large percentage of low-income, minority students.

Population: Participants for this study include high school juniors who attend low-income, mostly minority schools. Researchers plan to recruit a total of 6,400 students across two cohorts. Of these students, 1,600 will receive V-SOURCE, 2,400 will receive Milestones, and 2,400 will serve as the control group. The programs are designed for students who academically qualify to attend a 4-year college, so participants must, minimally, be on track to qualify for admission to a California State University campus.

Intervention: The original SOURCE program was a 15-month program that includes a college-application curriculum and one-on-one college advising focused on helping students with college selection, essay writing, application completion, financial aid, and deciding among college acceptances. It cost about $1,000 per student. V-SOURCE was developed to address lessons learned from SOURCE and in addition it also includes: (1) a SAT curriculum for students scoring below the national average on any of the three sections of the SAT; (2) small financial incentives for students upon completion of important milestones in the college application process (e.g., registering and taking the SAT or ACT, submitting applications to colleges, and applying to FAFSA); and (3) high school students have advisor-advisee interactions with a team of college student advisers through virtual technologies (e.g., phone, text, email, MySpace, Facebook, instant message, and Skype). As a result, V-SOURCE is expected to reduce the cost to about $400 per student. Under Milestones, students receive automated reminders, basic instructions, and small financial incentives to complete the key college application steps on time at a cost of about $140 per student.

Control Condition: Students in the control condition will employ a "business-as-usual" approach with access to resources that are normally available to high school students applying for college (e.g., access to college counselors and college centers at their high schools).

Research Design and Methods: An experimental design will be used to determine whether V-SOURCE increases college enrollment compared to both the control condition and the Milestones program, and whether Milestones increases college enrollment compared to the control. Approximately 6,400 high school juniors (two cohorts of about 80 juniors at each of 40 schools) who have expressed an interest in taking part in the study and who are eligible to take part will be randomly assigned to either the V-SOURCE program, the Milestones program, or to the control group. Randomization will include blocking on key variables for which subgroup analyses will be done: gender, parents' first language, and whether parents attended college. Students will be followed through their expected first year of college.

Key Measures: Key outcome variables include enrollment in 4-year college and persistence through the first year of college and are to be obtained from the National Student Clearinghouse. A student survey will be used to collect information on students' demographic and academic backgrounds, plans for the future, educational expectations and expectations for returns to education, and understanding of the college application process. Additional student information (e.g., standardized test scores, GPA, and eligibility for attending a 4-year CA public college or university) will be collected from district data. Intermediate outcomes (mediators) on SAT registration, SAT taking, Free Application for Federal Student Aid (FAFSA) submission, college application and acceptance, and college selectivity will be collected through surveys and administrative data.

Data Analytic Strategy: Researchers will estimate regressions of key outcomes and intermediate outcomes on indicators for each treatment group using both fixed effects models (using a school fixed effect) and random effects model (using a two-level model of students in school). For binary outcome variables, probit or logit models will be used. Subgroup analyses will be done using interaction terms between the treatment indicator and the subgroup variables. Fidelity of implementation will be examined according to five criteria (adherence; exposure; quality of delivery; participant responsiveness; and program differentiation) using advisor logs, observations and interviews, technology records, and a survey of high school seniors.