Measuring latent well-being with the PERMA-Profiler: A Multidimensional Item Response Theory (MIRT) Analysis

The PERMA-Profiler is a validated self-report instrument designed to measure psychological well-being, including PERMA (Positive Emotions, Engagement, Relationships, Meaning, and Accom-plishment). Despite consistent validation of its descriptive utility, and latent five-factor structure of PERMA using Confirmatory Factor Analyses (CFA), debates persist regarding the distinctiveness and incremental validity of these dimensions. This study applied MIRT to assess the between-item multidimensionality of the PERMA-Profiler and to evaluate its 11-point response scale with Category Boundary Discrimination (CBD) parameters. Data were drawn from a culturally diverse sample (N = 2,337), including previously published datasets. Results showed that a Rasch-compliant multidimensional Partial Credit Model did not fit the data, questioning the scientific validity of mean or sum scores. The PERMA Model did not fit due to stochastic de-pendencies in Meaning and unmodeled latent associations, likely caused by high latent correlations and convergence issues. Dependencies among Meaning items suggested that item M1 may need revision to focus on a specific aspect of meaningfulness rather than a global evaluation. Two-dimensional models, excluding item E3, showed adequate fit when using the multidimen-sional Graded Response Model. However, their latent association and broad theoretical scope raise questions about their incremental validity. Analysis of the 11-point scale revealed participants predominantly used the upper end, with categories 5 through 10 accounting for 84% of responses. Categories from 0 to 3 showed minimal discrimination, suggesting redundancy. Reducing the number of response categories, and using more objective descriptors, especially in Engagement, may enhance overall measurement precision.