STEM Workshops and Students’ Interest in Mathematics, Physics, and Computer Science: Machine Learning Approach

Abstract

Although STEM education is present in many countries and various aspects of this concept are being extensively researched, many educational systems still rely heavily on subject-specific learning. This study aims to examine the attitudes of students aged 14-17 regarding whether their participation in STEM workshops integrating content from mathematics, physics, and computer science contributes to an increased interest in studying these individual subjects. More specifically, we sought to determine whether the degree of increased interest in learning these STEM subjects could be predicted based on a set of dependent variables, including students’ sociodemographic data and their attitudes toward the different aspects of the conducted STEM workshop. The data was analyzed using 18 ML classifiers, with outlier removal methods applied to the five models that yielded the best results. The best-performing model was the decision tree with the IF automatic outlier removing technique, achieving an accuracy of 0.94. The key factors contributing to students’ increased interest in learning mathematics, physics, and computer science were primarily level of their engagement in the STEM workshop, beliefs about newly acquired knowledge, age, prior experience with STEM workshops, and current grade in physics.