Conclusions from meta-analytic structural equation models generally do not change due to corrections for study...

Area of Science:

• Psychometrics
• Quantitative Psychology
• Meta-analysis

Background:

• Meta-analytic structural equations modeling (MASEM) is a powerful tool for theory testing.
• Debate exists on whether to use raw or artifact-corrected correlations in MASEM.
• Artifacts like measurement unreliability can influence meta-analytic results.

Purpose of the Study:

• To investigate the impact of artifact corrections on structural equations modeling (SEM) fit indices.
• To examine the effect of artifact corrections on the stability of paths within SEM.
• To determine if corrections for study artifacts alter substantive conclusions in MASEM.

Main Methods:

• Comparison of SEM fit indices using mean observed correlations versus artifact-corrected correlations.
• Assessment of path stability (significance, magnitude, rank order) under both correction methods.
• Analysis conducted on meta-analyzed correlation matrices common in psychological research.

Main Results:

• Model fit indices were generally similar regardless of whether correlations were artifact-corrected.
• Path stability, including significance and relative strengths, remained largely unaffected by corrections.
• Substantive conclusions drawn from the structural models were consistent across correction approaches.

Conclusions:

• Artifact corrections in MASEM typically do not alter overall model fit or path stability.
• Findings suggest that standard corrections for study artifacts are generally safe for theory testing in psychology.
• Substantive conclusions are robust to artifact correction methods when variable reliability is typical.