Abstract: This study showcases how genetic algorithms can be used to derive a short screening scale out of longer assessment instruments, in cultural adaptations. Leveraging a sample of 1,081 participants aged 6–18 years collected with the Conners-3 Parent version in Romania, we employed a genetic algorithm to identify an optimal item subset that maximizes screening accuracy for ADHD. The adaptation process balanced psychometric rigor with cultural sensitivity, comparing the newly derived short scales to the original C3AI. Validation analyses, including confirmatory factor analysis (CFA), reliability estimates (Cronbach’s α and McDonald’s ω), and receiver operating characteristic (ROC) analysis, demonstrated strong diagnostic power, acceptable reliability, and structural validity. Notably, the Romanian-adapted scale exhibited improvements in model fit and diagnostic utility compared to the original short screening form. This research highlights the potential of automated methodologies, such as genetic algorithms, to enhance cultural adaptations of psychological measures while retaining psychometric integrity. The findings underscore the utility of context-specific scale redevelopment to optimize clinical screening tools.
Iliescu et al. (Tue,) studied this question.