The test in the field of education is a measuring tool used by educators and to support the process of assessing the abilities of students. However, all items contained in the instrument need to be tested. This study aims to test the construct validity of the Cognitive Abilities (CA) questionnaire instrument with the Computer-Based Test (CBT) model. The research sample of this study consists of 31 students. The method used to collect data is a questionnaire. The instrument used in collecting data is a questionnaire. The assessment of the questionnaire instrument was carried out by five experts who were invited as validators in the process of the Focus Group Discussion (FGD) technique. Checking the construct validity of the questionnaire was carried out by confirmatory factor analysis (CFA) using LISREL 9.30, with a small sample. The result confirmed that the value of approximate fit (chi-square) = 0.00 ≤ 0.03, degree of freedom (df) = 0 ≤ 2.00, the Root Mean Square Error of Approximation (RMSEA) = 0.000 ≤ 0.08 and p-value 1.000.000 > 0.80. Based on the results, the data shows a better fit with the correlated one-factor model or single item or questionnaire (content, construction, and language style). The point of view that construct validity emphasizes more on how far the questionnaire compiled is related to the theoretical measurement of the concepts that have been prepared by asking for expert judgment.
Copyrights © 2021