Poster Paper: Remedial Screening Tests, High School Grades, and College Success

Thursday, November 7, 2013
West End Ballroom A (Washington Marriott)

*Names in bold indicate Presenter

Zun Tang and Sarah Truelsch, City University of New York
Community colleges have traditionally administered standardized screening (basic skills) tests to identify students who have remedial needs to successfully complete college-level coursework. However, recent studies (Scott-Clayton, Crosta, & Belfield, 2012; Martorell & McFarlin, 2011; Calcago & Long, 2008) challenged the validity of this test-only approach of remediation policies. The results point to the fact that an alternative placement approach using multiple measures, specifically high school transcript information in addition to standardized tests, is likely to increase placement accuracy and reduce mis-assignment errors without compromising success rates in college-level courses.

Using rich administrative data from a large urban public community college system, we examined the predictive power of remedial screening tests and high school grades on college success. We use both OLS and logistic regression models to estimate models predicting GPA in college courses and passing or failing a course.

From our initial analysis of one entering cohort (N=2,326), we find that high school grades are better indicators of later college success (grades in gateway math courses) than basic skills standardized exams alone. Results are robust to a number of alternative definitions of criterion math courses for students in STEM and non-STEM majors.

With this empirical grounding, we build a placement index that combines standardized test scores and high school grades to create a metric that improves remediation placement, defined as maintaining or improving overall grades in criterion courses while decreasing the number of remedial placements. We examine (simulate) the potential impact of implementing this placement index by examining the sensitivity and specificity measures from the logistic model, which allows us to evaluate the performance of our new remediation placement policy.