Evaluating a Peer Assessment Approach in Introductory Programming Courses

Abstract

This paper describes a study of how peer assessment can be introduced to novice programmers to support programming knowledge. In the field of computer science, peer assessment is not a common approach in beginner programming courses as novice programmers generally may not have enough programming knowledge to assess their peers effectively. Therefore, this study contributes in empowering novice programmers and improving their learning quality at early stages. Thus, this study's aims are threefold: firstly, to identify an appropriate category-based assessment method, secondly, to determine how a peer assessment framework can be considered useful and valid as a tutor's assessment may be, and thirdly, to measure the students’ performance after a peer assessment activity. To achieve the first aim, a pilot study was completed across two universities with 42 undergraduate computer science participants. To accomplish the second and third aims, an experimental study with 170 undergraduate students was conducted. The results indicated that both rubric and marking scheme forms could be used to support peer assessment, as there was no significant difference in the results of use found between them. Moreover, although the convergence level of peer assessment with tutor assessment was not aligned, it was at a medium level with the students’ average scores when compared with the tutors’ assessments. In addition, the practice of peer assessment was found to have a positive effect on subsequent achievement for those students who took part in the mock peer assessment exercise.