Structural validity and measurement invariance of the Japanese version of the Kessler Psychological Distress Scale (K6): an item response theory analysis

Abstract

Abstract Background The Kessler Psychological Distress Scale (K6) is in widespread use to screen for psychological distress, but evidence to support the quality of the Japanese version (JK6) remains limited, in particular with respect to item response theory (IRT)-based item parameters and the test information function (TIF). Evidence concerning measurement invariance across gender and age groups likewise remains limited. Therefore, this study examined the structural validity and measurement invariance of the JK6 in an online community sample of 6632 Japanese-speaking adults aged 20–59 years. Materials and methods Confirmatory factor analysis (CFA) was used to compare one- and two-factor models. Using IRT with a graded response model, item discrimination and threshold parameters were estimated and item information functions and the TIF were derived. Measurement invariance across gender and age groups was tested using multi-group CFA (MCFA), and differential item functioning (DIF) was examined via ordinal logistic regression. Results The CFA results supported a primarily unidimensional interpretation of the JK6. IRT analyses showed very high item discrimination, and the TIF indicated high measurement precision for mild-to-moderate levels of psychological distress. MCFA supported scalar invariance across gender and age groups, and the DIF effect sizes were negligible. Conclusions These findings indicate that the JK6 is a primarily unidimensional measure of psychological distress that provided precise and comparable assessment across gender and age groups in an online community sample of Japanese-speaking adults aged 20–59 years. Further studies should be undertaken to examine the generalizability of these findings across the broader Japanese adult population, including older adults and clinical populations.