Association for Public Policy Analysis & Management

Students from low-income backgrounds face substantial barriers to college entry and success. Prior research demonstrates that interventions focused on providing students with information about college and financial aid and access to assistance can increase college enrollment (e.g., Bettinger et al, 2012; Hoxby and Turner, 2013), and text messaging has been shown to be a cost-effective approach to increase students’ college and financial literacy and to connect them to professional support when they need help (Castleman & Page, 2015, 2016).

We investigate the efficacy of a text messaging campaign designed to provide high school students with timely and digestible information all along the pathway from college exploration during junior year through to college matriculation in the fall after high school graduation. We conducted this investigation with two distinct samples. First, in collaboration with the College Board and uAspire, both national non-profit organizations, we implemented text-message based outreach and advising to students in over 700 US high schools that primarily serve low-income students. From these high schools, students gave consent to be involved in the program when they took the PSAT as high school juniors in October 2014. Based on this recruitment method, we delivered text-based outreach to approximately 70,000 high school students in the class of 2016 to provide important college and financial aid information, with the goal of improving their postsecondary opportunities and outcomes. Second, we collaborated with several Texas-based school districts to implement a version of the text-based intervention. The Texas sample primarily was made up of urban districts in the Houston and Austin metropolitan areas. All students in the high school graduating class of 2016 for whom the participating schools had cell phone numbers received outreach.

For both samples, the key experimental intervention involved text-message based outreach on a regular basis to remind students about key steps in the college search, application, selection and transition process and to provide them guidance and personalized support in navigating these processes. Within the “national” sample, the messaging was centralized, with all treatment group students receiving the same message content, and the messages being sent on behalf of the College Board by advisors who were employed as full-time at uAspire (headquartered in Boston, MA). The text-outreach to treatment-assigned students invited text-based, two-way communication with a designated advisor. Within the Texas sample, messages were generally sent by students’ high school counselors, and as such, message content varied modestly across participating Texas high schools to align with the context of each school. In both samples, outreach began in Spring 2015 and continued through September 2016.

We used a cluster randomized control trial design. For both samples, we conducted randomization at the school level. We rely on several sources of data that allow us to examine impacts on key steps in the college-going process as well as initial college enrollment and early patterns of college persistence. We are currently analyzing data from both experiments and will be prepared to share comprehensive experimental results and associated conclusions / implications.