Checking the inventory: Illustrating different methods for individual participant data meta-analytic structural...

Area of Science:

• Psychological Methods
• Statistical Modeling
• Meta-Analysis

Background:

• Researchers often have access to raw data for meta-analyses.
• Structural Equation Modeling (SEM) is a powerful tool for analyzing complex relationships.

Purpose of the Study:

• To identify, illustrate, and compare SEM analysis options when raw data are available.
• To discuss differences in procedures, capabilities, and outcomes of various SEM techniques.

Main Methods:

• Directly analyzing raw data using multilevel and multigroup SEM.
• Using summary statistics with correlation-based meta-analytical SEM (MASEM).
• Fitting a path model based on the theory of planned behavior to multiple datasets using open-source software.

Main Results:

• Differences observed in parameter estimates and standard errors across methods.
• Variations in handling missing data, including study-level moderators, and conceptualizing heterogeneity.
• Direct data analysis and summary statistic-based approaches yield distinct outcomes.

Conclusions:

• Applied researchers need clear guidelines for conducting individual participant data MASEM.
• Further research is essential to establish best practices for meta-analytical SEM with raw data.