Latent variable mixture modeling in psychiatric research--a review and application.

Area of Science:

• Psychology
• Psychiatry
• Statistical Modeling

Background:

• Population heterogeneity is a key challenge in psychological and psychiatric research.
• Latent variable mixture modeling offers a powerful framework for understanding this heterogeneity.
• Existing methods may not fully capture the complex structures of psychological phenomena.

Purpose of the Study:

• To provide a non-technical introduction to latent variable mixture modeling.
• To illustrate its application in a cross-sectional context.
• To explore the latent structure of psychotic experiences.

Main Methods:

• Latent Class Analysis (LCA) for subgroup classification.
• Factor Mixture Modeling (FMM) combining LCA and factor analysis.
• Application to data on psychotic experiences.

Main Results:

• Latent variable mixture models effectively identify homogeneous subgroups.
• Factor mixture models accommodate both categorical and dimensional data.
• These methods are adaptable to complex psychiatric and psychological phenomena.

Conclusions:

• Latent variable mixture modeling is a flexible tool for studying population heterogeneity.
• It enables researchers to compare dimensional, categorical, and hybrid construct conceptualizations.
• This approach is valuable for complex phenomena in psychology and psychiatry.