Measurement invariance of hierarchical taxonomy of psychopathology across sociodemographic covariates: A moderated nonlinear factor analysis approach.

The Hierarchical Taxonomy of Psychopathology (HiTOP) provides a dimensional alternative to categorical nosology. Using a nationally representative sample (N = 16,238), we first replicated the hierarchical structure via bass-ackward exploratory factor analysis. Moderated nonlinear factor analysis (MNLFA) was used to test differential item functioning, mean impact, and factor score differences for HiTOP dimensions across categorical (i.e., sex assigned at birth and race/ethnicity) and continuous (i.e., age and income) sociodemographic covariates and their two-way interactions. Results for MNLFA revealed sources of measurement noninvariance that were not previously detected using multigroup confirmatory factor analysis. Both significant diagnostic differential item functioning and latent factor mean-level differences emerged at the spectrum and subfactor levels across sociodemographic moderators. In contrast to previous studies supporting robust measurement invariance of the HiTOP structure, the results of the present study demonstrate that the HiTOP latent dimensions are generalizable only once measurement noninvariance, especially for categorical diagnostic data, is identified and adjusted for. These results also highlight the unique information provided by the MNLFA approach in assessing the generalizability of HiTOP. (PsycInfo Database Record (c) 2026 APA, all rights reserved).