A Categorical Principal Component Regression on Computer-Assisted Instruction in Probability Domain

2018-01-01
The purpose of this study is to examine the effects of computer-assisted instructional material (CAIM) prepared in R program on eighth grade students’ permutation–combination and probability achievement and their attitudes toward computer-assisted learning. In the study, we collect survey data from 74 conveniently selected students and their schools; data consists of 45 highly correlated explanatory variables with different measurement levels. To deal with the multicollinearity problem among mixed type of explanatory variables, first, we apply categorical principal components analysis (CATPCA), and hence, reduce the dimension of data. In the following, we use uncorrelated components instead of the original correlated variables to fit the multiple linear regression (MLR) model in order to question whether CAIM has been effective in teaching probability domain. Results show that the general success of the students and basic socioeconomic and technological factors affecting this situation and interaction of those factors with the secondary social situation of the student’s family have statistically significant effects on the probability achievement of the students. Instruction method is also found as a statistically significant factor in explaining students’ achievement in permutation–combination subjects. However, none of the explanatory variables considered in the study are found statistically significant in explaining the attitudes of students toward computer-assisted instruction (CAI).
Citation Formats
T. Kapucu, Ö. İlk Dağ, and İ. Batmaz, A Categorical Principal Component Regression on Computer-Assisted Instruction in Probability Domain. 2018, p. 164.