Bayesian and Maximum-Likelihood Modeling and Higher-Level Scores of Interpersonal Problems With Circumplex Structure

Area of Science:

• Psychology
• Psychometrics

Background:

• Interpersonal behavior difficulties are commonly assessed using the circumplex-based Inventory of Interpersonal Problems (IIP).
• The IIP's eight scales can be structured into three factors: Dominance, Love, and Distress.

Purpose of the Study:

• To evaluate the higher-level structure of interpersonal problems using Bayesian confirmatory factor analysis.
• To compare Bayesian and maximum likelihood confirmatory factor analysis for assessing circumplex structures.
• To validate the IIP's three-factor model using external criteria.

Main Methods:

• Employed Bayesian confirmatory factor analysis and maximum likelihood confirmatory factor analysis.
• Utilized data from a non-clinical online questionnaire study (N = 822).
• Included external criteria: Big Five traits (Agreeableness, Extraversion, Neuroticism) and grandiose narcissism.

Main Results:

• Successfully replicated the three-factor structure of the IIP using both methods.
• Observed high proximity of Bayesian loadings to perfect circumplexity.
• Demonstrated excellent reliability (≥0.90) and validity of higher-level scores for Dominance, Love, and Distress.

Conclusions:

• Bayesian confirmatory factor analysis is well-suited for analyzing circumplex structures in interpersonal problems.
• The three-factor model of the IIP is valid and reliable.
• Recommends using higher-level scores for Dominance, Love, and Distress as parsimonious, reliable, and valid measures.