Affiliations: [a]
University of Michigan, Ann Arbor, MI, USA
| [b]
University of Virginia, Charlottesville, VA, USA
| [c]
Virginia Commonwealth University, Richmond, VA, USA
Correspondence:
[*]
Address for correspondence: Paul B. Perrin, Ph.D., Professor of Data Science and Psychology, Elson Student Health Center, University of Virginia, 400 Brandon Ave., Charlottesville, VA 22903, USA. E-mail: [email protected].
Abstract: BACKGROUND: Despite the widespread international use of the Affiliate Stigma Scale in the context of disability, much research on its psychometric properties has taken dubious statistical approaches. OBJECTIVE: The aim of this study was to examine the psychometric properties of the Affiliate Stigma Scale in a sample of Turkish spinal cord injury/disorder (SCI/D) caregivers. METHODS: Participants completed the Affiliate Stigma Scale, Zarit Burden Interview, and Generalized Anxiety Disorder-7. This study conducted a series of confirmatory factor analyses (CFAs) of the Turkish Affiliate Stigma Scale using its originally theorized 3-factor structure and a 1-factor structure, as well as an exploratory factor analysis (EFA) to refine the scale items to create a short form, with a final CFA of the items in the short form. RESULTS: The 3-factor and 1-factor CFAs of the 22 items from the Affiliate Stigma Scale suggested poor fit to the data across every fit index. An EFA yielded four factors, although the patterns of item loading onto the factors did not map in any discernible way to the original subscales theorized by the scale creators. A 1-factor CFA with the six items loading onto factor 1 of the EFA showed much better fit indices, with most achieving good or adequate fit. CONCLUSION: The current study supports a 1-factor solution with a short form comprised of six items, at least in Turkish and with SCI/D caregivers.
