Skip to main content
ABCT 2025 59th Annual Convention Main banner
Psychology Tools Ad
Final Program


Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.
Presentation Icons
Recorded Session
Recorded Session
Ticketed Session
Ticketed Session

Addictive Behaviors

(PS1-26) Substance use disorder dimensionality redux: Revisiting the appropriateness of DSM-5's unidimensional model of substance use disorder

Friday, November 21, 2025
10:00 AM - 11:00 AM CST
Location: Elite Hall A, Level 1
Abstract Text: Since the DSM-5, substance use disorders (SUDs) have been conceptualized as unidimensional, with abuse and dependence collapsed into one construct. A unidimensional model was justified empirically through the use of latent variable and item response theory modeling, with these models showing strong fit. Nevertheless, simulation studies have documented that, when considering categorical and especially binary data, conventional fit indices like scaled TLI, CFI, and RMSEA consistently overestimate model fit quality. In their place, robust fit indices for categorical data (Savalei, 2020) were developed and shown to gauge model fit with considerably improved validity. Here, we re-evaluate the (uni)dimensionality of the DSM-5 SUD criterion set using data from the population-based National Survey on Drug Use and Health 2021-23 (N = 165,249). We fit 1-factor models across eleven substances (e.g., alcohol, opioids, cannabis, sedatives) and extracted conventional (scaled) and robust fit indices. Scaled and robust estimates diverged sharply, with the conventional indices supporting a unidimensional model as hypothesized (i.e., TLI and CFI > 0.95). By contrast, no unidimensional models were supported by robust fit indices. Exploration of the dimensionality of SUD across 11 substances corroborated multidimensionality, suggesting 2-factor models that parsed substance consumption and problems, for 5 of the 11 substances (i.e., cannabis, hallucinogens, inhalants, prescription drugs, stimulants). Our results highlight the importance of assessing construct dimensionality with models and statistical indices that are appropriate to the data in-question, especially when determining field-standard questionnaires for use in both clinical and research settings. Further, the differing dimensionality across substances calls into question the appropriateness of a unified criterion set for all substances and raises the possibility of overdiagnosis on the basis of consumption in absence of dysfunction, or vice versa. The presentation thus appeals broadly to those interested in SUDs as well as those concerned with classification and psychometrics in general.