Evidence for a unidimensional 15-item version of the CASP-19 using a Rasch model approach

摘要

Aim: The aim of this study is to explore the dimensionality of the CASP-19 - a quality-of-life measure for early old age - and to assess the stability of the item locations over time and changes in person measures. Methods: The Rasch partial credit model was applied to the CASP-19 data obtained from Waves 1 and 3 (N = 9,407 and N = 7,732) of the English Longitudinal Study of Ageing to assess item fit, dimensionality and differential item functioning (DIF). Results: Five items demonstrated category disordering for the last two response categories of the CASP-19 (which were collapsed), and 4 items demonstrated item misfit and were removed from the analysis. Cronbach's alpha indicated good internal consistency for both waves (0.87 for Wave 1 and 0.88 for Wave 3). There was no evidence of multidimensionality, and no significant DIF was observed for gender. When the two datasets were combined, no statistically significant differences were observed between the two sets of item locations. A statistically significant difference was found for the mean person measure [t(5,524) = 17.88, p<0.001)], suggesting quality of life improves as people age. Conclusion: A 15-item unidimensional factor structure was observed for the CASP-19, which was stable over time.