The paucity of underrepresented minorities (URMs) earning science, technology, engineering, and mathematics (STEM) degrees remains an issue in revitalizing the U.S. biomedical workforce. Due to reductions in federal funding, maintaining the integrity of programs that focus on URM retention and recruitment is crucial. We present data on the mechanisms used to recruit URM students to our program (e.g., email, events, referrals, website), which individually were equally effective in attracting applicants to the program. Recruitment mechanisms were grouped and further classified relative to their cost to implement as lower and higher cost. Our results indicate that lower cost mechanisms, statistically, were as effective as higher cost mechanisms in recruiting students who persisted to PhD programs. Using a binary logistic regression model to predict PhD matriculation, higher cost mechanisms were not significant predictors of PhD matriculation. Collectively, these data demonstrate for the first time that lower cost mechanisms can be as successful in recruiting URM students to summer programs who pursue PhDs in STEM fields.

Science, technology, engineering, and mathematics (STEM) diversity research programs seek to make progress in increasing the number of underrepresented students that pursue STEM at the highest degree levels. Yet few programs have outlined their path to help students achieve the STEM PhD. Our program, Opportunities in Genomics Research (OGR), showed significant increases in PhD matriculation over 8 years of National Institutes of Health (NIH) funding. We explored typical measures, which include grade point average (GPA), institution classification, and graduate school ranking, and found that these measures alone do not explain the improved outcomes. We examined changes in the application materials as proxies for commitment to a PhD degree. These data show a significant correlation of desired degree pursuit to increased proxies and proxy type (open-or close-ended questions answered by applicant or referee). We demonstrate that changes in application procedures for diversity research programs correlate to improved program outcomes with statistical significance.