Meta-analysis of Monte Carlo simulations examining class enumeration accuracy with mixture models.

Area of Science:

• Social and Behavioral Sciences
• Quantitative Psychology
• Statistical Modeling

Background:

• Mixture modeling is increasingly utilized in social and behavioral sciences.
• Monte Carlo simulation studies are crucial for evaluating statistical methods.
• Synthesizing findings from simulation studies is essential for methodological advancement.

Purpose of the Study:

• To provide a methodological guide for conducting meta-analyses of Monte Carlo simulation studies.
• To investigate simulation design factors influencing class enumeration accuracy in mixture modeling.
• To inform methodologists on planning future simulation studies.

Main Methods:

• A meta-analysis framework was applied to selected Monte Carlo simulation studies on mixture modeling.
• Data were analyzed using generalized linear mixed models and meta-regression.
• Key steps included literature identification, screening, data extraction, analysis, and interpretation.

Main Results:

• Different fit indices demonstrate varying performance based on sample size.
• Bayesian information criterion (BIC) accuracy increases with larger sample sizes.
• Entropy performance is superior in conditions with smaller sample sizes.

Conclusions:

• Meta-analysis of simulation studies offers valuable insights into methodological factors.
• Findings guide the selection of appropriate fit indices based on study conditions.
• This approach enhances the reliability and applicability of simulation research in mixture modeling.