Artificial intelligence literacy scale: A study of reliability and validity in Turkish university students

Abstract

This study aims to adapt to Turkish the "Scale for the assessment of non-experts: AI literacy" developed by Laupichler et al. (2023a). The scale consists of 31 items with three sub-dimensions: technical understanding, critical thinking, and practical applications. The data required for the validity and reliability study of the scale were collected from 642 undergraduate and graduate students studying in different departments of a state university in the fall semester of the 2023-2024 academic year. First, CFA was applied to the data according to the factor structure in the original scale, but as acceptable fit values could not be obtained because of the analysis, exploratory factor analysis was performed. In the reliability analysis of the factor structure determined by EFA, KMO was calculated as =0.948. It was determined that the scale items were collected in three factors and explained 61.1% of the total variance ("critical thinking" is 25.8%, "technical knowledge" is 25.2%, and "practical applications" explains 10.2% of the total variance). As a result of EFA, it was seen that the sub-dimensions of some of the items in the original scale had changed, and since the factor load values of the three items were very close to each other, they were removed from the scale. Because of CFA, which was conducted to evaluate whether the data supported the hypothesized relationships between the measured variables, Cronbach’s alpha value was found to be 0.90. As a result of the CFA analysis conducted with the 3 sub-dimensions and 28 items in the scale, the Chi-square value (X²=2.85; df=345, N=317, p< .001), which is the fit index of the model, has a good fit and is significant, SRMR = 0.0545and RMSEA = 0.077 values and fit indices and the model has an acceptable fit.

Ähnliche Arbeiten