Turkish Adaptation and Psychometric Evaluation of the Colorado Learning Attitudes About Science Survey (CLASS) in Physics

Abstract

The Colorado Learning Attitudes about Science Survey (CLASS) is an instrument to measure student beliefs about physics and learning physics. In this research, Turkish adaptation and psychometric evaluation of the CLASS is discussed. In the first stage, the translation process, which included examination of six experts (four experts in physics education and two experts in English and Turkish languages) for content validity and 13 student interviews for face validity, was described. In the second stage, exploratory (EFA) and confirmatory (CFA) factor analysis results obtained from 1391 freshman students were discussed for construct validty. The EFA yielded three factors that consisted of 20 items, which explained 39.61 % of the total variance. These factors were named as: Problem Solving Effort, Conceptual Understanding, and Personal Interest and Real-World Connection. Based on the CFA results, the three-factor 20-item instrument showed acceptable fit statistics. Compared to the original CLASS, the proposed version with 20-item model was shorter, easier to administer and easier to score, valid and reliable survey, so feasible to use. The overlapping of the constructs (i.e., violation of the unidimensionality) in the original CLASS was a problem, while in the proposed model none of the items were included in more than one construct.