Matching Strategies for Observational Studies with Multilevel Data

July 30, 2012

Peter Steiner has received support from the Institute of Education Sciences to develop a theoretical background for matching designs and techniques with observational multilevel data and to derive guidelines for educational researchers who have already started using multilevel matching techniques in their analyses. The project will investigate different within- and across-cluster matching strategies for estimating causal treatment effects across and within clusters. Steiner pursues two goals in using theoretical investigations, simulation studies, within-study comparisons, and analyses of ECLS–K data. First, to determine under which conditions within- and across-cluster matching strategies produce consistent estimates of average treatment effects. Second, to find which matching strategies and analytic approaches work best in educational research practice. The guidelines Steiner develops should contribute to the implementation of more and better-warranted matching designs in the future.