Psychometric Assessment of the Bangla Version of General Health Questionnaire-12 (GHQ-12)

Abstract

Abstract: The Bangla-translated 12-item General Health Questionnaire (GHQ-12) has been widely used in Bangladesh, but no attempt has been made to assess its psychometric properties. We investigated the latent structure, item quality, and differential item functioning of the Bangla GHQ-12 among 788 Bangladeshi adults (197 clinical outpatients and 591 nonclinical university students). A split-half exploratory factor analysis supported a unidimensional structure, confirmed by a bifactor model with the best fit (CFI = .96; TLI = .99; SRMR = .02; RMSEA = .06) in confirmatory factor analysis. The scale demonstrated excellent reliability (ordinal α = .96; McDonald’s ωt = .97) and known-group validity (clinical vs. nonclinical subgroups, t = 290.21, p < .001; Cohen’s d = 1.46). Item response theory-based analysis indicated reliability and coverage across the general mental health continuum (θ = −3 to 3). While no item bias was found across genders, two items were more sensitive to clinical outpatients. We recommend the Bangla GHQ-12 as a unified tool for assessing general mental health and psychological distress among Bangladeshi nationals.

Keywords

• Bangla GHQ-12
• item response theory
• structural validity
• differential item functioning
• mental health
• Bangladesh