Year of Publication



Martin School of Public Policy and Administration

Executive Summary

High rates of poverty and low levels of educational attainment have plagued the Appalachian region throughout history. The Robinson Scholars Program was created in 1996 as part of the University of Kentucky Board of Trustees’ plan to support economic and community development in Appalachian Kentucky.

The Robinson Scholars Program is a scholarship and student support program that serves first-generation college and college-bound students from twenty-nine Eastern Kentucky counties with historically low rates of college attendance. The Program’s mission is to empower students to complete baccalaureate degrees and thereby add to the educational capital of their communities. The first class of Robinson Scholars was selected as eighth-graders in 1997. Since its inception, the Program has named approximately 540 students as Robinson Scholars.

The Robinson Scholars Program likely affects not only the Scholars themselves, but also their families and communities. The visibility of Robinson Scholars in their communities, including those who have graduated from the University of Kentucky and returned to the area to pursue their careers, may inspire others in the community to pursue higher education as well. The purpose of this study is to investigate this phenomenon by collecting and analyzing empirical data to determine if such a “peer effect” or “role model effect” does, in fact, exist. The existence of an effect will highlight indirect and unintended benefits of the Program in addition to the benefits enjoyed by the Scholars themselves. The following research questions were explored:

  • Did implementation of the Robinson Scholars Program have the effect of increasing high school graduation rates in the 29-county service area?
  • Did implementation of the Robinson Scholars Program have the effect of increasing college matriculation rates in the 29-county service area?

To answer these questions, a panel data set for 171 of Kentucky’s 176 school districts was used first to estimate two simple linear regression models and then to estimate two fixed effects models. The dependent variables in the models were district high school graduation rate and college matriculation rate in school year t, for t from 1994-1995 to 2004-2005. The fixed effects models were determined to be preferred to the simple linear regression models.

The results of the fixed effects models show no statistically significant effect of the Robinson Scholars Program on high school graduation and college matriculation rates in the service area. Though there is no evidence that the Program is having an effect, the results suggest that the model estimation could be improved by using individual level data. Such a data set could potentially control for omitted variables, which may be biasing the coefficients of the models presented here.