The validation, reliability, and measurement invariance of the cognitive fusion questionnaire in Chinese community-dwelling adults

Abstract

Background: There was no study on the validation and measurement invariance test of the Cognitive Fusion Questionnaire (CFQ) among Chinese community-dwelling adults. This study aimed to evaluate the reliability, validity and measurement invariance of the 7-item CFQ for cognitive fusion among Chinese community-dwelling adults in Hong Kong. Methods: In total, 552 Chinese adults (18–64 years old; 51% males) in Hong Kong finished the online survey during March to April 2023. The questionnaire included the CFQ, the Acceptance and Action Questionnaire-II (psychological flexibility), the Kessler Psychological Distress Scale (psychological distress), the Short Warwick-Edinburgh Mental Wellbeing Scale (mental well-being), and the self-rated mental health. Confirmatory factor analysis (CFA) and multi-group CFA were conducted. Results: The unidimensional structure of the scale was confirmed by CFA, with a satisfactory fit to the sample (χ2(14)= 15.935, p = 0.317, CFI = 0.998, RMSEA = 0.024, SRMR = 0.021). The standardized factor loadings ranged from 0.71 to 0.78. Full scalar invariance across educational level and psychological distress level and partial scalar invariance across gender and age groups were confirmed by multi-group CFA. The CFQ showed good internal consistency (Cronbach’s α = 0.90), and convergent validity with significant correlations with experiential avoidance (rs(430) = 0.69, 95% CI 0.63 to 0.74), psychological distress (rs(430) = 0.61, 95% CI 0.55 to 0.67), mental wellbeing (rs(430) = -0.29, 95% CI -0.38 to -0.20) and self-rated mental health (rs(430) = -0.24, 95% CI -0.33 to -0.15), with p-values all < 0.001. Conclusions: This study is the first to examine the psychometric properties of the 7-item CFQ in Chinese community-dwelling adults in Hong Kong. The CFQ is a brief, reliable and valid measure for assessing CF among Chinese community-dwelling adults. Caution is warranted for direct comparison across samples with uneven age and gender distributions.