Going “Up” to Move Forward: S-1 Bifactor Models and the Study of Neurocognitive Abilities in Psychopathology

Abstract

Cognitive abilities have been theorized as important to the development and maintenance of a wide range of disorders. However, inconsistent associations between cognitive abilities and specific disorders have meant understanding any functional mechanisms are difficult. The primary reliance of the nosological approach to diagnosis may contribute to the heterogeneity. Recently, structural models of psychopathology have dominated much of the literature and given rise to the p factor. These models combat the diagnostic stability and comorbidity issues of the nosological approach. However, recent evidence argues against p being a substantive construct, limiting conclusions that can be drawn between p and cognitive abilities. Here, we argue for the use of the S-1 bifactor approach, where the general factor is defined by cognitive abilities, to explore the association. We use simulation techniques to give show how S-1 bifactor models can be used to examine the relationship, and how the results can be interpreted.

Meta-analytic tests of measurement invariance of internalizing and externalizing psychopathology across common methodological characteristics.

“…The difference in the covariation of Internalizing and Externalizing between sample types was more substantial than the difference in factor loadings (although still only a moderately sized effect) and suggested-consistent with prior work (e.g., Haywood et al, 2021; Note. MDD = major depressive disorder; RMSEA = root mean squared of approximation; SRMR = standardized root mean-square residual; CFI = comparative fit index; DCFI = CFI difference between nested models; U = latent variable variance/covariance; INT = Internalizing; EXT = Externalizing; ASPD = antisocial personality disorder; GAD = generalized anxiety disorder; // = 0.…”

One p-Factor for All? Exploring the Applicability of Structural Models of Psychopathology within Subgroups of a Population

“…Recently, authors have described the S−1 bifactor approach to the study of psychopathology as an alternative to the standard bifactor approach [ 24 , 25 , 26 ]. The S−1 approach allows for the predefinition of the substantive construct of the general factor by loading theoretically important indicators representing a domain of interest directly on the general factor [ 26 , 28 ]. Indeed, the S−1 approach combats many of the issues that we present here with regard to the bifactor model.…”