Assessment of English–French differential item functioning of the Satisfaction with Appearance Scale (SWAP) in systemic sclerosis. (September 2017)

• Record Type: Journal Article
• Authors: Jewett, Lisa R.; Kwakkenbos, Linda; Hudson, Marie; Baron, Murray; Thombs, Brett D.
• Is Part Of: Body image. Volume 22(2017)
• Page Start: 97
• Page End: 102
• Publication Date: 2017-09
• Subjects: Satisfaction with Appearance Scale -- Systemic sclerosis -- Body image -- Measurement -- Differential item functioning; Body image -- Periodicals; Body image -- Research -- Periodicals; Body Image -- Periodicals; 306.4613
• DOI: 10.1016/j.bodyim.2017.06.005
• Languages: English
• ISSNs: 1740-1445

Abstract

Highlights: Significant body image concerns are common in systemic sclerosis (SSc). As a rare disease, studies in SSc often include patients from multiple countries. The Satisfaction with Appearance Scale (SWAP) lacks cross-linguistic validation. SWAP scores can be combined for English- and French-speaking SSc patients. Abstract: The Satisfaction with Appearance Scale (SWAP) has been used to assess body image distress among people with the rare and disfiguring disease systemic sclerosis (SSc); however, it has not been validated across different languages groups. The objective was to examine differential item functioning of the SWAP among 856 Canadian English- or French-speaking SSc patients. Confirmatory factor analysis was used to evaluate the SWAP two-factor structure (Dissatisfaction with Appearance and Social Discomfort). The Multiple-Indicator Multiple-Cause model was utilized to assess differential item functioning. Results revealed that the established two-factor model of the SWAP demonstrated relatively good fit. Statistically significant, but small-magnitude differential item functioning was found for three SWAP items based on language; however, the cumulative effect on SWAP scores was negligible. Findings provided empirical evidence that SWAP scores from Canadian English- and French-speaking patients can be compared and pooled without concern that measurement differences may substantially influence results.