Psychometric properties of a virtual science labs assessment questionnaire using confirmatory factor analysis and neural networks among middle school students

Abstract

This study aimed to investigate the psychometric properties of a questionnaire assessing virtual science labs among middle school students. It also sought to uncover the difficulties faced by these students in utilizing virtual labs for science education, examine the factorial structure of the assessment questionnaire, estimate its reliability, and validate its construct validity using confirmatory factor analysis (CFA). Additionally, the study aimed to diagnose the structure of the concept of awareness about the importance of virtual labs through neural network analysis. To achieve these goals, the study employed a descriptive correlational methodology. The study instrument consisted of a questionnaire with 33 items distributed across four dimensions: awareness, participation, obstacles, and suggestions. The questionnaire was administered to a sample of 97 female students. Results from exploratory factor analysis (EFA) revealed four factors with high loading values. Confirmatory factor analysis indicated a good fit for the four-factor model with the data (RMSEA=0.091, CFI=0.941, TLI=0.939). In light of the neural network analysis indicators, the most influential items affecting the others were W7, D2, D6, P5, and P7. The internal consistency reliability (Omega squared) for the overall questionnaire was 0.957, and the alpha coefficient for the overall questionnaire was 0.932, indicating a high level of reliability and validity. Additionally, the results showed no statistically significant differences between males and females in the virtual labs assessment questionnaire.