A customized and automated assignment management and marking system for evaluating student performance in the STEM disciplines

CONTEXT A strong increase in student numbers for CIV3204, an introduction to statistics unit taught at Monash University within an undergraduate engineering course, has been accompanied by decreased performance in the final exam. Anecdotal evidence suggests that this is caused by cheating in the in-semester assignments. Based on evidence in the literature that individualized assignments result in reduced cheating, and that automated marking allows for the completion of more assignments by the students (leading to more practice and feedback), a system to generate and automatically mark individualized assignments has been developed. A closed-form solution does not exist for most questions, so existing methods such as Moodle quizzes could not be used. This paper provides an overview of the system and the very positive results of the implementation. PURPOSE OR GOAL The objective of this study is to improve the students' performance in the final exam for CIV320 and their learning in the unit. The hypothesis is that automatically marked individualized assignments lead to reduced cheating and the completion of more assignments, and consequently an improved performance in the final exam. A user-friendly system working through Moodle has been developed for this purpose. APPROACH OR METHODOLOGY/METHODS An individualized assignment was generated for each of the 11 topics in the unit, which was automatically marked. Detailed feedback was provided to the students afterwards. This level of assessment would have been impossible to achieve with manual marking. The performance of the cohort in the final exam and the Student Evaluation of Teaching and Units (SETU) are used to evaluate the system. ACTUAL OR ANTICIPATED OUTCOMES The system has led to very positive results. 66% of the students appreciated that the assignments were the most effective aspects of the unit. The unit received its highest overall satisfaction in eight years. The failure rate in the final exam decreased from 22% in 2019 to 11% in 2020, even though the final exam was more difficult. CONCLUSIONS/RECOMMENDATIONS/SUMMARY The greatest surprise from the study was that the students were very positive about the large number of assignments, and the automated marking. The students suggested to improve the python-based Graphical User Interface system, which we will replace with a website. The system improved the students' learning through more practice and feedback, evidenced by their achievement in the exam. Based on the positive outcome, we suggest that automated marking should be further developed and implemented in the STEM disciplines.