Construct Validity of Computer-Based Physics Test (CBPsyT) Through Confirmatory and Item Response Theory: A Complementary Approach to Test Development
Construct Validity of Computer-Based Physics Test (CBPsyT) Through Confirmatory and Item Response Theory: A Complementary Approach to Test Development

Summary/Abstract: Background/purpose. This study aims to validate the computer-based physics test (CBPsyT) construct through confirmatory factor analysis (CFA) to ensure that CBPsyT is suitable for its intended use. However, CFA studies are based on weak theoretical perspectives, a lack of testing of alternative theoretical views, or insufficient evidence for construct validity, so item response theory (IRT) analysis needs to be added to this study. Materials/methods. The research is exploratory and includes a quantitative approach. We developed CBPsyT according to test development procedures consisting of twelve steps and administered it to 516 students. Construct validity analysis based on the CFA and IRT perspectives using the R program. Results. The relationship between indicators and latent constructs, with multiple representations in CBPsyT, is indicated by indicator loadings> 0.70, which denote a strong relationship between the two. Internal consistency is demonstrated by a composite reliability > 0.80, and CFA model fit values indicate that the hypothesized model is consistent with the empirical data. The IRT results add that discrimination and difficulty scores range from -2 to +2. The test information function indicated that the instrument was reliable for ability (Theta) scores ranging from -3 to +3. Conclusion. The test quality profile was complete when CFA and IRT were combined. To the best of our knowledge, this article provides practical information on the psychometric characteristics of CBPsyT and guides new researchers in demonstrating construct validity.

• Details
• Contents