Gokhan AKSU1, Cigdem REYHANLIOGLU KECEOGLU2
1Aydın Adnan Menderes University, TURKEY.
2Gaziantep College Foundation Private Schools, TURKEY.
DOI: 10.14689/ejer.2019.84.6
ABSTRACT
Purpose: In this study, Logistic Regression (LR), CHAID (Chi-squared Automatic Interaction Detection) analysis and data mining methods are used to investigate the variables that predict the mathematics success of the students.
Research Methods: In this study, a quantitative research design was employed during the data collection and the analysis phases.
Findings: The findings obtained in this study showed that the variables, which had significant effects on the mathematical success of the students in each method, differ from each other. Although the independent variables with a significant effect on the dependent variable were different according to different methods, the findings indicated that the importance order of the variables did not change according to the method used. In this study, the correct classification ratios obtained by the class concerning PISA mathematics literacy differed by different methods.
Implications for Research and Practice: CHAID analysis and REPTree algorithm may be an alternative for one another in the studies that aimed to classify individuals concerning their success. However, LR analysis should not be considered as an alternative method since it will provide significantly different results compared to the other two methods.
Keywords: CHAID Analysis, Logistic Regression Analysis, Data Mining, PISA.