Comparing structural models for internalizing pathology: Latent dimensions, classes, or a mix of both?

Area of Science:

• Psychology
• Psychopathology
• Quantitative Psychology

Background:

• Internalizing problems conceptualization evolved from categorical to dimensional and hybrid models.
• Most research focuses on disorder-level analysis, with limited symptom-level examination of internalizing constructs.
• Hybrid approaches integrating categorical and dimensional aspects are under-explored at the symptom level.

Purpose of the Study:

• To compare the structural and concurrent validity of categorical, dimensional, and hybrid models for internalizing constructs at the symptom level.
• To evaluate model fit and predictive power across diverse samples (adults, students, patients).
• To determine the optimal approach for assessing internalizing symptoms and their relationship with external variables.

Main Methods:

• Latent class analysis (categorical), confirmatory factor analysis (dimensional), and semi-parametric factor analysis (hybrid) were employed.
• Data from four samples (N=2455) using the Inventory of Depression and Anxiety Symptoms-II were analyzed.
• Concurrent validity was assessed using measures of disability, externalizing symptoms, personality, and quality of life in subsamples.

Main Results:

• Dimensional models demonstrated superior structural validity and concurrent predictive power compared to categorical and hybrid models.
• Models allowing for non-normal distributions within the dimensional framework showed the best performance.
• Dimensional models consistently explained a significant portion of the variance in external variables (median adjR² = .16–.18).

Conclusions:

• Dimensional models are recommended for future research on internalizing constructs due to their superior validity and predictive utility.
• Adopting dimensional approaches facilitates empirical integration in clinical science and enhances personalized mental health assessment.
• This approach can lead to more precise identification of individual needs and guide more effective interventions.