Predicting Students Performance Through Behavior and Computational Thinking in Programming

Abstract

The advancement and achievement of the students are known as primary success metrics for many universities around the world. Hence, universities invest heavily in resources to reduce the performance gap between academically good performers and low-performing students. The high rate of failure in computer programming courses can attribute several factors. This paper explores the educational history of students, their fields of study, and their approaches to learning applied to programming courses. Considering students’ attendance, assignment marks, program practice at home, willingness to work, interaction in class predicts student’s behavior. By considering students’ lab test marks, willingness to apply a new algorithm predicts student’s computational thinking. The data are collected from undergraduate students by conducting lab tests and surveys among teachers. Finally, considering student behavioral and computational thinking parameters, and by using simple linear regression, the attributes which are supported and which are not supported for predicting student performance can be finalized.