A flexible moderated factor analysis approach to test for measurement invariance across a continuous variable

Area of Science:

• Psychometrics
• Statistical Modeling

Background:

• Moderated factor models test measurement invariance across continuous variables.
• These models traditionally require specifying a parametric functional form for the moderator, which can be challenging.

Purpose of the Study:

• To present a semiparametric moderated factor modeling approach.
• To overcome the limitations of assuming a specific functional form in moderated factor models.

Main Methods:

• Developed a semiparametric moderated factor model.
• Evaluated the model's performance through a simulation study.
• Applied the model to an intelligence dataset.

Main Results:

• The semiparametric model demonstrated viability in parameter recovery.
• The model showed adequate power for distinguishing measurement invariance models.
• Successful application to a real-world intelligence dataset.

Conclusions:

• The semiparametric approach offers a flexible alternative for moderated factor analysis.
• This method is effective for measurement invariance testing when the functional form is unknown.
• The model is a valuable tool for analyzing complex psychological data.