Analysis on reliability and validity of SF-36 scale in urban residents

Abstract

To evaluate the reliability and validity of SF-36 scale in urban residents, and provide reference for the selection of suitable health measure tools for urban residents. Multi-stage cluster stratified sampling was conducted to select the residents aged ≥18 years in three urbanized communities of Hangzhou. SF-36 scale was used for the measurement of the quality of life and Spearman-Brown and Cronbach' s α coefficients were used for the evaluation of split-half reliability and internal consistency reliability. The convergent and discriminative validity were evaluated by using the success rate of experiments and the criterion-related validity was evaluated with correlation analysis and non-parameter test. Structural equation modeling was used in the evaluation of contract validity. SF-36 scale had good split-half reliability (R=0.94) and internal-consistency reliability (except for bodily pain and vitality, Cronbach's α range: 0.70-0.91). The convergent validity (88.57%), discriminate validity (successful rates 90.61%) and the criterion-related validity (γs=0.56, the score was consistent with the self-reported health status) were good. Second-order confirmatory factor analysis model was not well-fitted (GFI= 0.721, AGFI= 0.682, CFI= 0.731, RMR= 0.084, RMSEA= 0.098), indicating that the construct validity was poor. The reliability, consolidation validity, discrimination validity and criterion-related validity of SF-36 scale were good, while the construct validity was poor. Improvement is needed when the scale is used for urban residents.