1. Measurement invariance of the University of Rhode Island Change Assessment Scale in Project MATCH: An exploratory structural equation modeling approach
- Author
-
Dylan K. Richards, Frank J. Schwebel, Kiefer D. Cowie, Matthew R. Pearson, and Katie Witkiewitz
- Subjects
Adult ,Male ,Psychiatry and Mental health ,Alcohol Drinking ,Cognitive Behavioral Therapy ,Psychometrics ,Latent Class Analysis ,Humans ,Medicine (miscellaneous) ,Female ,Factor Analysis, Statistical ,Toxicology ,Article - Abstract
BACKGROUND: Progression through the stages of change is proposed to be a mechanism underlying the effects of treatment for alcohol use disorder (AUD). However, examining stages of change as a mechanism of treatment effects requires that the measure is invariant across patient subgroups, treatment conditions, and time. In the present study, we examined measurement invariance of the University of Rhode Island Change Assessment Scale (URICA) in Project MATCH using an exploratory structural equation modeling (ESEM) approach. METHODS: We conducted a secondary data analysis of Project MATCH (N=1726; M(age)=40.2, SD=10.9; 75.7% male; 80.0% non-Hispanic white), a multisite randomized clinical trial that tested three AUD treatments: Motivational Enhancement Therapy, Cognitive-Behavioral Therapy, or Twelve-Step Facilitation. Participants completed the 24-item URICA for assessing the stages of change in relation to drinking at baseline and posttreatment (3 months after baseline). RESULTS: A 4-factor ESEM provided a good fit to the data and a better fit to the data than a conventional 4-factor confirmatory factor analysis model. Further, the URICA demonstrated scalar invariance across each patient subgroup at baseline (sex, ethnicity, marital status, education, and parental history of AUD) and treatment condition at follow-up. However, the URICA was not longitudinally invariant as the metric model resulted in a significant decrement in model fit. CONCLUSIONS: The measurement invariance of the URICA over time was not supported. Longitudinally invariant measures of the stages of change are needed to appropriately test the proposal that progression through the stages explains treatment effects.
- Published
- 2022