Title

Correcting For The Distorting Effects Of Study Artifacts In Meta-Analysis

Abstract

The methods presented in this chapter for correcting meta-analysis results for sampling error and biases in individual study results might appear complicated, but the reader should be aware that more elaborated and extended descriptions of these procedures are presented in Hunter and Schmidt (2004) and that Windows-based software (Schmidt and Le 2004) is available to apply these methods in meta-analysis. It is also important to bear in mind that without these corrections, for biases and sampling error the results of meta-analysis do not estimate the construct-level population relationships that are of greatest scientifi c and theoretical interest (Rubin 1990; Schmidt and Hunter 2004, chapters 1, 14). Hence, in meta-analysis, these corrections are essential to the development of valid cumulative knowledge about relations between constructs. This is especially important in light of recent developments concerning the use of meta-analytic results in the testing of causal models. Meta-analytic results are increasingly being used as input to path analyses and other causal modeling techniques. Without appropriate corrections for measurement error, range restriction, and other artifacts that bias meta-analysis results, these causal modeling procedures will produce erroneous results. Copyright © 2009 by Russell Sage Foundation. All rights reserved.

Publication Date

12-1-2009

Publication Title

The Hand. of Res. Synthesis and Meta-Analysis, 2nd Ed.

Number of Pages

317-333

Document Type

Article; Book Chapter

Personal Identifier

scopus

Socpus ID

84902733577 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/84902733577

This document is currently not available here.

Share

COinS