Methods for synthesizing regression models: The case of education production functions

Loyola University Of Chicago, Chicago IL

Investigators

Abstract

This proposal seeks to develop and test new methods for meta-analysis and synthesis of research that uses multiple regression models of different specificity. Current methods for meta-analysis focus on the synthesis of effect sizes based on summary statistics such as standardized mean differences, correlations and odds ratios. The intellectual merit of this proposal is the development of methods for synthesizing the results of multiple regression across studies. The methods proposed in this study build on statistical methods for missing data.