The comparison of several factorial structures of the Cornell Critical Thinking Test Level Z

Abstract

The Cornell Critical Thinking Test (CCTT) level Z is a broadly applied instrument for the measurement of undergraduate, gifted students’ and adults’ critical thinking abilities, however, there is a lack of studies in the literature investigating its factorial structure. The test developers emphasize the interpretation of the test results as a general critical thinking factor, including different cognitive abilities, but also highlight the overlap between the items of the subfactors. The aim of the study is the investigation of internal consistency, the comparison of different factorial structures (unidimensional, correlated, and hierarchical models), and gender invariance testing of the CCTT level Z. Hungarian-speaking undergraduate students participated in the study (N = 825). For the confirmatory factor analysis (CFA) the Mplus version 8.7, with weighted least squares mean and variance adjusted estimation was used. Measurement invariance of the test across genders was analyzed using Muthén’s two-step procedure for dichotomous data. The results indicated that the correlated four-factor and second-order structures of the test exceeded the acceptable model fit criteria. Post hoc inspection was conducted on the second-order four-factor model, indicating a shortened 22-item version of the test (with a general critical thinking factor, including four subfactors: deduction, meaning and fallacies, induction, assumption identification) with excellent fit indices [χ2 (203) = 259.309, CFI = .967, TLI = .963, RMSEA = .018, SRMR = .056]. Configural and scalar invariance of the abbreviated Hungarian version of the test across genders were confirmed.

Keywords

• critical thinking
• undergraduate students
• confirmatory factor analysis
• measurement invariance