Predicting middle school students’ mathematics achievement using decision trees

Abstract

In today's world, it is crucial to raise individuals who can think critically and flexibly, see things from a variety of perspectives, and develop creative solutions to common challenges in many countries including Türkiye. One of the ways to attain this purpose is to provide qualified mathematics education. During this qualified education determining the extent to which students meet the program goals is an important and fundamental part of the process of educating the individuals. Education should be provided in accordance with the requirements of the age using a curriculum with qualified measurement, assessment, and evaluation practices. This study aims to predict middle school students’ achievement at four grade levels in mathematics at the end of the semester utilized as a predictor measurement and assessment approach. Chi-squared Automatic Interaction Detection (CHAID), a decision tree analysis, was used to examine the importance of exams, quizzes, homework, and teacher observation in predicting the end-of-semester achievements. As a result of the study, when the first semester is examined in general, quizzes are mostly determined as the best predictor of student success, while exams are generally for the second semester. Implications for teachers and students are stated.