The Short Warwick‐Edinburgh Mental Wellbeing Scale in Swedish: A Large‐Scale Study of Reliability and Validity in Adults
Frida Ekbom Lodin, Johnny PellasABSTRACT
Background and Aims
The Short Warwick‐Edinburgh Mental Wellbeing Scale, SWEMWBS, is a self‐rating scale designed to measure positive aspects of mental health. The aim of the present study was to investigate the psychometric properties of the official Swedish translation of the SWEMWBS within a representative sample of Swedish adults.
Methods
Data from 31,802 participants aged 18–102 years were retrieved from the public health survey Life and health. Reliability was investigated by examining the internal consistency with McDonald's omega in different age groups. Convergent validity was evaluated by examining Spearman's rho correlations with the six‐item Kessler psychological distress scale (Kessler‐6). Structural validity was investigated through exploratory and confirmatory factor analysis, and measurement invariance across age groups was examined using multi‐group confirmatory factor analysis.
Results
The internal consistency was strong, with omega values ranging from 0.87 to 0.89 in the different age groups. The correlations between SWEMWBS and Kessler‐6 in the different age groups were statistically significant and strong, ranging from −0.62 to −0.75. The factor analyses offered support to the single‐dimension structure but showed better fit for a two‐factor model. Measurement invariance testing demonstrated stable psychometric properties across age groups.
Conclusions
The official Swedish translation of SWEMWBS has strong reliability and convergent validity. Although the two‐factor model showed superior fit, a strong correlation between factors indicated substantial overlap suggesting that the scale could be considered unidimensional, in line with previous research. The difference in factor structure compared to the original English version might be due to language differences or the gathering of data during the covid‐pandemic. The most noticeable strength of the current study is the large age‐stratified sample, including the oldest old.