Comparisons of Artifact Correction Procedures for Meta-Analysis: An Empirical Examination on Correcting Reliabilities
Date of Award
Doctor of Philosophy (PhD)
This study reviewed some challenges and issues in artifact correction meta-analysis, particularly around using reliability estimates to correct for measurement error. Two individual correction procedures—the Hunter-Schmidt procedure and the procedure developed by Raju, Burke, Normand, and Langlois (the RBNL procedure)—are addressed in this research. The purpose of this study is to use real-world data to examine the differences between meta-analytic estimations produced by the two artifact correction procedures and those by the traditional bare-bones meta-analysis procedures, under the condition of inter-dependent reliabilities. The impact of this inter-correlation on meta-analysis results needs investigation when artifact indicators, such as reliability of predictor and reliability of outcome, are proven to be significantly inter-correlated. The current study revealed that neither the choice of artifact correction nor the choice of analysis procedure provided any significant differences in the estimation results, whereas it was the choice of the reliability estimates that generated noticeable differences in the results. In addition, the violation of the assumption for independent reliability did not greatly impact the meta-analytic estimation results.
Zhao, Lei, "Comparisons of Artifact Correction Procedures for Meta-Analysis: An Empirical Examination on Correcting Reliabilities" (2017). Dissertations. 2605.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.
Copyright © 2017 Lei Zhao