Croon's Bias-Corrected Estimation for Multilevel Structural Equation Models with Non-Normal Indicators and Model...

Area of Science:

• Educational research
• Quantitative psychology
• Statistical modeling

Background:

• Multilevel structural equation models (MSEMs) are crucial for analyzing complex educational data with latent variables.
• Croon's bias-corrected factor score (BCFS) path estimation offers a promising approach for MSEMs, especially with limited sample sizes common in educational research.
• The performance of BCFS in MSEMs under challenging conditions like non-normal indicators and model misspecification requires thorough investigation.

Purpose of the Study:

• To evaluate the accuracy and efficiency of BCFS estimation for MSEMs.
• To assess BCFS performance under conditions of non-normal indicators and model misspecifications.
• To compare BCFS with other estimation methods for MSEMs in educational research contexts.

Main Methods:

• Conducted two simulation studies to assess BCFS estimation in MSEMs.
• Varied conditions included limited sample sizes, non-normal indicators, and model misspecifications.
• Compared the accuracy and efficiency of BCFS against other estimation techniques.

Main Results:

• BCFS estimation for MSEMs demonstrated greater dependability, efficiency, and reduced bias compared to other methods under limited sample sizes and model misspecifications.
• BCFS estimation was found to be more susceptible to non-normality in indicators.
• The findings support the utility of BCFS as an alternative or supplemental estimator for MSEMs.

Conclusions:

• BCFS estimation is a valuable tool for multilevel structural equation modeling in educational research, particularly when dealing with constrained sample sizes or model misspecifications.
• Researchers should consider potential impacts of non-normal indicators when employing BCFS.
• The study encourages the use of BCFS as a robust estimation strategy for complex educational research designs.