Alternative Approaches to Validate a Critical Thinking Measurement Instrument for University Students
Alternative Approaches to Validate a Critical Thinking Measurement Instrument for University Students

Summary/Abstract: Background/purpose. The aim of this study is to present classical and robust alternatives for the validation of an instrument that assesses critical thinking (CP). The Chilean adaptation of the Halpern Critical Thinking Assessment Using Everyday Situations (HCTAES) test was analyzed using three factorial models and a short-version model, with two estimation methods: maximum likelihood (ML) and weighted least squares (WLSMV). Internal consistency was examined with point and interval estimation for different reliability coefficients. Materials/methods. The sample consisted of 214 students from a Colombian university. The instrument used was the Chilean adaptation of the Halpern Critical Thinking Assessment Using Everyday Situations (HCTAES), which assesses five dimensions of critical thinking through open-ended and closed-ended items. For the psychometric analysis, three-factor models and an abbreviated version were considered, using Maximum Likelihood Estimation (MLE) and Weighted Least Squares (WLS) as estimation methods. Internal consistency was evaluated using reliability coefficients with confidence intervals, while construct validity was examined using Confirmatory Factor Analysis (CFA), complemented by convergent and discriminant validity indicators. Results. The results indicate that the complete test and the open-ended question factor exhibit satisfactory internal consistency (ω=0.78 and ω=0.79, respectively). The confirmatory factor analysis using the robust WLSMV estimator yielded better fit across the absolute, incremental, and parsimony fit indices, corroborating Halpern's theoretical model. Conclusion. It is concluded that the configural and open-format models retain adequate psychometric properties, i.e., good fit and stable factor configurations across the evaluated structures. The study contributes methodological elements to the line of validation and evidence of psychometric properties for PC, or serves as a reference for other analogous constructs.

• Details
• Contents