Pedagogical Analysis of Correlation Between Students’ Expenditure of Time for Laboratory Sessions and Exam Score Using Statistical Tools

Abstract

This study examines the correlation between time spent in mandatory laboratory sessions and the result on the final exam. By conducting this study we hope to find ways to identify different students’ needs, such that deep learning can be stimulated. The data from 141 students have been collected during a single course session in an undergraduate course at Lund University.

Using a Gaussian mixture model, the data points could be clustered such that we could identify four different clusters of students:
1. Medium-to-fast students with high exam score,
2. Students using quite long time at the laboratory, with high exam score,
3. A linear cluster where the exam result increase as they stay longer at the laboratory sessions, and
4. A group of students with lower exam scores, and who stay long at the laboratory sessions.
Different ideas are discussed for the different groups, with the common goal to stimulate the students’ deep learning. Such ideas can be to identify students that are falling behind during the laboratory session, or ensuring that the students really understand the preparatory assignments before they start the laboratory session.