Psychometric Properties of the Turkish Version of the ChatGPT Usage Scale

Abstract

Artificial Intelligence (AI) tools are actively used by academics and students. ChatGPT stands out among the AI tools most used by undergraduate and graduate students. ChatGPT offers students idea development and academic writing support for their assignments and research. This study examined the validity and reliability of the ChatGPT Usage Scale in Turkish culture. The ChatGPT Usage Scale consists of three subscales: Academic Writing Aid, Academic Task Support, and Reliance and Trust. The study participants consisted of 332 undergraduate and graduate students. Reliability analyses conducted on the ChatGPT Usage Scale show that Cronbach's alpha and McDonald's omega values ranged from .87 to .96. According to the results of the study's Confirmatory Factor Analysis (CFA), the fit indices of the ChatGPT Usage Scale were determined to be good. The results of the scale's item discriminant analysis and item-total correlation values were found to be at acceptable levels. Furthermore, the scale's high item factor loadings and convergent validity demonstrate its construct validity. All research findings demonstrate that the Turkish version of the ChatGPT Usage Scale is reliable and valid. Therefore, this scale is expected to make significant contributions to future studies on the integration of AI and ChatGPT into the Turkish education system.

Ähnliche Arbeiten