Designing A Prototype Model of Peer Assessment for Introductory Computer Programming Courses

Abstract

This thesis aims to provide an in-depth, contextual understanding of how peer assessments can be integrated, as a learning process, within introductory programming courses. It was motivated by exploring several difficulties first-year programmers encounter in introductory programming courses. One main issue is the difficulty of getting timely feedback from teachers. With a large class, a teacher may not be able to give instant feedback. Furthermore, students often lack the confidence to complete individual tasks, because they have not yet developed an effective internal model of a computer that they can use to construct viable knowledge. This study focuses on peer assessment and its effectiveness since peers can serve as a valuable source of instant feedback. Additionally, interactions with peers can increase students' confidence and improve their problem-solving abilities. Students and teachers can be reluctant to use peer assessment, as programming assignments are highly practical and require sufficient knowledge to complete. To identify students' receptiveness to peer assessment, as well as teachers’ attitudes to implementing such an activity, this study adopted a mixed-methods approach. Statistical analyses revealed that participants were generally positive about engaging in formative peer assessment in programming courses, but they differed in some areas which is why the validity of peer assessment among students, as well as the impact of peer assessment on their performance were also examined. The results indicate that, at a moderately medium level, first-year assessors and teachers are similar in assessment; moreover, peer assessment exerts a positive influence on student performance in programming skills. The literature has not clarified whether programming students have specific needs regarding peer assessment. A qualitative analysis of students’ expectations related to implementing peer assessment provided crucial details about students’ requirements in this regard. Students noted that some elements encourage them to use peer assessment, such as clear rubrics, self-assessment, rewards for their efforts and visual feedback. Visual feedback, particularly for either author or reviewer, is an unfamiliar aspect in the context of peer assessment; hence, it is one focus of this study. Students are mainly concerned with the credibility of the reviewers giving feedback on their work. A Balanced Allocation algorithm has been developed to retrieve a group of reviewers to assess the work of each author, with a view to show students feedback of better quality. The peer programmer prototype website is the output that contains a group of requirements that could be considered when developing a peer assessment for programming students. This type of study is invaluable for teachers who are concerned with peer assessment for students, as it informs practice and provides guidance; thus, teachers can relate this research to their own contexts.