Abstract / Summary
Purpose: Semantic associative processing is essential for language, cognition, and communication and is frequently impaired in neurodegenerative disorders. However, no brief Turkish instrument currently assesses semantic associations across both abstract and concrete conceptual domains. This study aimed to develop the Turkish Semantic Association Test (SEMAT), establish normative data, and examine selected psychometric properties.
Method: The SEMAT was developed using stimuli representing abstract, biological, and manmade concepts. Face validity was evaluated through expert review and pilot testing. Normative data were obtained from 365 cognitively healthy Turkish-speaking adults aged 53-82 years. Demographic effects were examined using multiple regression analyses, with age, education, and sex entered simultaneously as predictors. Given the bounded, negatively skewed score distributions and substantial ceiling effects, empirical percentile-based norms were generated. Internal consistency and convergent validity were assessed using semantic measures from the Batterie d'Évaluation Cognitive du Langage (BECLA-Tr). Item-level difficulty and discrimination were also examined.
Findings: Educational attainment was the only demographic variable independently associated with performance across all SEMAT scores, supporting education-adjusted norms. The SEMAT demonstrated excellent internal consistency (Cronbach's α = 0.88; McDonald's ω = 0.92). Significant correlations were observed between SEMAT scores and the BECLA-Tr semantic composite (ρ = 0.41-0.54, all p < 0.001), supporting convergent validity. Item-level analyses showed generally high accuracy and positive but variable discrimination across items.
Conclusion: These findings indicate that the SEMAT is a brief, culturally appropriate, and psychometrically sound measure of semantic associative processing with preliminary evidence of reliability and convergent validity and with education-adjusted normative reference data for cognitively healthy middle-aged and older Turkish-speaking adults.