Development of Japanese version of King's Stigma Scale and its short version: Psychometric properties of a self-stigma measure.

Abstract

The study aimed to develop two Japanese versions of King's Stigma Scale, a full version (KSS-J-1) and a short version (KSS-J-2), through psychometric property testing. The sample included 112 people with mental illness. We tested the constructs of the scales using both confirmatory and exploratory factor analyses. Internal consistency and test-retest reliability were tested. We examined convergent validity with self-esteem or perceived stigma, and different group validity, using the Kessler Psychological Distress Scale (K6). Whereas a relatively weak model fit (comparative fit index = 0.66, Tucker-Lewis index = 0.63, root mean square error of approximation = 0.097) of KSS-J-1 (full version: 28 items) was found, KSS-J-2 (short version: 17 items), produced by exploratory factor analysis, had a moderate model fit (comparative fit index = 0.90, Tucker-Lewis index = 0.89, root mean square error of approximation = 0.063). High internal consistency (KSS-J-1, ω = 0.82-0.89; KSS-J-2, ω = 0.86-0.89) and moderate test-retest reliability (KSS-J-1, interclass correlation = 0.56-0.88; KSS-J-2, interclass correlation = 0.45-0.85) were reported. Some subscales and the entire scale of KSS-J-1 were significantly correlated with self-esteem and perceived stigma. Conversely, only two subscales in KSS-J-2 were significantly correlated with self-esteem. The scores of each subscale and the entire score for both KSS-J-1 and KSS-J-2 in the high psychological distress group were higher than the low group (KSS-J-1, d = 0.61-0.83; KSS-J-2, d = 0.47-0.70), except for the Discrimination subscale in KSS-J-2. Both Japanese versions of King's Stigma Scale can be utilized depending on their intended use.