The Effectiveness of Synthetic Control Methods for School Comparison in Program Evaluation

• In Event: Poster Session 9
  In Poster Session: Program Evaluation in Schools Poster Session

Abstract

Synthetic controls are an innovative matching method from the economics field beginning to appear in educational program evaluations. Unknown is whether they offer dependable comparisons when applied using academic achievement measures. Longitudinal data from two large school districts were used to examine effect size differences necessary for detection by these methods. Results across 16 specifications indicated that demographic measures provided no additional predictive power beyond pretreatment test scores. Further, inconsistent results across data sets suggest that a few schools’ atypical performance, even within a well-matched set of comparison schools, gave required effect sizes outside of the range of reasonable school growth. Further investigations are recommended to better establish the effectiveness of synthetic controls for school-level comparisons.

Author

• Clay S. Johnson