<![CDATA[Inappropriate dispensing of antipsychotic medications may cause preventable medication-related harm, yet validated instruments for assessing pharmacists’ knowledge and dispensing practices remain limited in Indonesia. This study evaluated the Rasch measurement properties of a questionnaire developed for Indonesian pharmacists. A cross-sectional survey was conducted among 642 pharmacists recruited through convenience sampling using professional pharmacy networks and online communication platforms. Participants completed a 22-item instrument comprising 12 dichotomous knowledge items and 10 five-category practice items. 22-item instrument comprising 12 dichotomous knowledge items and 10 five-category self-reported practice items. The two domains were analysed separately using the Rasch model, assessing person and item fit, unidimensionality based on principal component analysis of standardized residuals (with a first-contrast eigenvalue <2.0), reliability, separation, item-person targeting, and sex-related differential item functioning. Item fit was evaluated using Infit and Outfit mean-square (MNSQ) statistics, with values between 0.5 and 1.5 considered acceptable. Mean person and item infit and outfit mean-square values were close to 1.00 in both domains. Person reliability was 0.78 for knowledge and 0.86 for practice, item reliability was 1.00 in both domains, and Cronbach’s alpha values were 0.76 and 0.87, respectively. The Rasch measures explained 60.5% of the variance in knowledge and 69.5% in practice. Wright maps showed satisfactory central targeting, and no visually apparent moderate-to-large sex-related differential item functioning was identified. Overall, the instrument demonstrated satisfactory initial Rasch-based validity and reliability, although further validation in geographically balanced samples and comparison with observed dispensing behaviour are warranted.]]>
