Examination of the Connor-Davidson Resilience Scale 10 (CD-RISC-10) using the polytomous Rasch model.

Abstract

The Connor-Davidson Resilience Scale's (CD-RISC) 10 item variant has previously demonstrated acceptable psychometric properties of its test scores using traditional methods (e.g., confirmatory factor analysis), and concurrent validity with resilience-related outcomes, particularly in samples of younger adults. While alternative methods of examining the psychometric properties of the long-form CD-RISC exist in the literature, the short-form measure has unclear evidence of local item independence and a unidimensional structure, which are key assumptions of a polytomous Rasch model approach to examining the measure's psychometric properties. The current study employed a sample of young adult university students (n = 708, xage = 26.43 years (s = 7.77)) on their nursing practicum placements to examine the CD-RISC-10 against the polytomous Rasch measurement model criteria. The analyses suggested a seven-item variant of the CD-RISC-10 performed acceptably, and omitted issues with local item dependence and item misfit. Effect sizes of the standardized parameters estimated for the 7-item and original 10-item versions of the CD-RISC-10, when predicting compassion fatigue and compassion satisfaction, were small (s = -0.24, s = -0.23) and moderate (s = 0.48, s = 0.47) for the respective measures, which suggested similar efficacy when examining the test scores' concurrent validity. The shorter version of the CD-RISC-10 consequently demonstrated generally acceptable psychometric properties for its test scores, and remained a parsimonious approach to examining individual psychological resilience that will benefit from further development. (PsycInfo Database Record (c) 2021 APA, all rights reserved).