Journal of Education & Social Policy

A Propensity Score Matching with Multiple Imputation Approach: Studying the Effects of Remedial Math-Taking in College on Degree Attainment
Meghan Ann Clovis, Mido Chang

Abstract
The purpose of this study was to investigate whether multiple imputation procedures (MI) and propensity score matching (PSM) would improve the estimation of remedial math-taking on college-degree attainment. The primary motivation was to explore a robust research method in an area of educational research that suffers from the inability to conduct a randomized controlled design. Analyses were conducted on both matched and unmatched groups, as well as on 12 multiply imputed complete-case data sets. Logistic regression analyses, both with and without the inclusion of covariates, indicated that remedial math-taking in college was a significant negative predictor of degree attainment. These results were consistent across unmatched groups, matched groups, and all 12 multiply imputed data sets. Neither multiple imputation procedures nor propensity score matching resulted in a significant improvement in the ability to control for preexisting group differences.

Full Text: PDF