Association for Public Policy Analysis & Management

We present evidence on the effectiveness of one strategy to address barriers due to the lack of direct experience with college. We maintain that the concrete experience of visiting a college campus during middle school could increase one’s comfort level with and interest in college, thereby motivating students to engage in pro-college behaviors during high school that would increase their readiness for and likelihood of attending college. Specifically, for this presentation, we ask whether visiting a college campus multiple times during eighth grade changes student motivation, college going behaviors, and intentions to apply to college.

937 8^th grade students from 15 schools in northwest Arkansas participated in this study in the 2017-18 school year. 41% of participating students are students of color, 35% report not having a college-educated parent, and 44% have never visited a college campus. Students were randomly assigned within schools to one of two conditions: to receive an information packet about postsecondary options or to participate in multiple campus visits in addition to receiving postsecondary information.

Students randomized to the visits condition were brought to the University of Arkansas three times, where they participated in activities designed to introduce them to various aspects of campus life. Specifically, students participated in the following three visits, each lasting approximately 4 hours:

1. Workshop on college life and how to prepare for college; walking tour of campus.
2. Tour of on-campus housing; workshops with 2 of 12 participating departments.
3. Razorback baseball game or a campus scavenger hunt.

Participating students were surveyed twice, first at baseline and again post-intervention. Our survey instrument assessed college knowledge, attitudes towards college, grit, self-management, perspective taking, and demographic information. Our randomized design allowed us to assess the causal impact of on-campus experiences and information relative to just information on these constructs using simple linear regressions, controlling for student’s demographic information and taking into account the block nature of the randomization. We also perform an impact-on-treated analysis to account for student absences on campus visits. Standard errors are clustered to account for the nested data structure.