Causal Effect Evaluation of Personalized Reminder Strategies on Government Welfare Program Enrollment: A Propensity Score Matching Approach

Abstract

Government digital service platforms face persistent challenges achieving optimal enrollment rates for welfare programs including SNAP and Medicaid. This research develops a causal inference framework quantifying personalized reminder intervention effects on enrollment completion, addressing selection bias through Propensity Score Matching, temporal dynamics via Difference-in-Differences, and endogeneity through Instrumental Variables. Methodology validation uses simulated observational data incorporating realistic population heterogeneity and non-random treatment assignment. Results demonstrate personalized reminders achieve 14.3 percentage point enrollment increases (p<0.001) after controlling confounding, with heterogeneous effects across age and digital literacy. The framework provides evidence-based guidance for optimizing government platforms per Executive Order 14058 on customer experience modernization.