Association for Public Policy Analysis & Management

Over the last decade, school districts around the nation have shifted their reform efforts from increasing high school graduation rates to increasing the number of students ready for college. As the New York City Department of Education (DOE) has shifted their focus to college and career readiness, it is important that they develop valid indicators with which to assess and track student progress. However, there is growing evidence that college readiness is not a one-size-fits-all concept. This study builds on prior literature by investigating the predictive validity of grades, test scores, high school coursework, and other factors on college outcomes with a particular focus on how predictive validity could vary across subgroups, including race/ethnicity, gender, language status, and socio-economic status.

Our study employs an extensive longitudinal database of students from The New York City Partnership for College Readiness and Success. This comprehensive database allow us to track DOE students as they move from high school to and through college in the city’s higher education system. To address our research question, we use a series of logistic regressions to identify the student engagement and achievement factors appropriate for inclusion in a college-readiness indicator for proximal and longitudinal outcomes (e.g., freshman grades and college completion). After developing our system-wide college-readiness indicators, we investigate how the probability of being identified as “college ready” fluctuates for various subpopulations of students.

Preliminary results suggest that the traditional measures of academic performance used to predict on-time high school graduation are not as reliable in predicting college outcomes, particularly for groups of students that are traditionally underrepresented in higher education. Ultimately, these results demonstrate the importance of going beyond the practice of using only test score cutoffs to define college readiness. Further, the lack of stability in the predictive ability of high school measures across student subgroups suggest the need for further development and collection of information on other potential factors that also contribute to post-secondary success, such as socio-emotional skills, college knowledge, and financial need.

Findings from this study can inform the efforts of of adminstrators and staff the secondary and post-secondary level. As K-12 systems across the U.S. work to build college readiness indicator systems to support school improvement and as higher education instituations use college readiness indicators in their admission or remediation placement processes, it is important to know whether these systems work similarly for all students and the potential tradeoffs in determining who is ‘college ready’ across various types of information used.