Since SAT and ACT tests have been long-suspected and then shown to contain class and race biases while not accurately predicting retention, the Schedler Honors College at the University of Central Arkansas (UCA) shifted to a holistic, multi-criterion selection process, de-emphasizing standardized tests, and then analyzed the outcomes. This essay discusses the statistical analysis of the admissions criteria at UCA, which served two goals: (1) test whether variables in the admissions model, developed in 2007, predicted retention--the results led to changes in the weighting of variables for a revised rubric used since 2010; and (2) improve enrollment of a more racially diverse population of students. Findings demonstrated that most variables used in typical higher education admissions protocols did not accurately predict retention in the Schedler Honors College at UCA. Only one variable correlated to retention in honors, namely, high school grade point average (hsGPA). By increasing the value of hsGPA in the revamped selection rubric, UCA was able to increase rates of retention as well as diversity of incoming students.