New Developments in Factor Score Regression: Fit Indices and a Model Comparison Test.

Area of Science:

• Statistics
• Psychometrics
• Econometrics

Background:

• Factor score regression (FSR) is widely used in structural equation modeling (SEM).
• Standard FSR methods can introduce bias in regression coefficient estimators.
• Existing bias correction methods exist, but further improvements in model fit assessment and inference are needed.

Purpose of the Study:

• To propose novel fit indices for FSR to evaluate model fit.
• To introduce a model comparison test for inference on regression coefficients within FSR.
• To compare the performance of FSR with bias corrections and SEM.

Main Methods:

• Development of new fit indices specifically for FSR.
• Introduction of a model comparison test utilizing these fit indices.
• A simulation study to evaluate bias, Type I error rate, and power.

Main Results:

• The proposed fit indices allow for inspection of model fit in FSR.
• The new model comparison test facilitates inference on regression coefficients.
• Simulation results provide a comparison of FSR, Croon's corrections, and SEM.

Conclusions:

• The proposed methods enhance the utility of FSR by addressing bias and improving model evaluation.
• These advancements contribute to more robust statistical inference in SEM contexts.
• The study provides valuable insights for researchers using FSR.