Comparison of Student-Level and School-Level Data in a National Impact Evaluation

• In Event: Poster Session 7
  In Poster Session: Division D Section 2 Poster Session

Abstract

The ability to understand what works in education requires rigorous research methods which often rely on student-level data which have become increasingly difficult to obtain. In this paper we explore the methodological alternative of using publicly available school-level data instead of student-level data to conduct an impact analysis of a national Expanded Learning Time (ELT) initiative on middle school students’ test scores. Multiple regression models were used to examine the effect of ELT on student achievement, using both student-level and school-level data. Overall, results from the two models were highly similar and conclusions would be the same whether based on student or school-level data. Effects on power, point estimates, and standard errors are discussed.

Authors

• Melissa Velez, Abt Associates Inc.

• Sarah Sahni, Abt Associates Inc.

• Alyssa Rulf Fountain, Abt Associates Inc.