Self and peer evaluations of student performance in skills based design projects

Self and peer assessment have been shown to help improve student engagement in assessment tasks, to improve behaviour and maintain interest and attention levels, and ultimately to improve student performance. In this study, we critically evaluated the process, reliability, validity, benefits and drawbacks of self and peer assessment in a first-year product design project with the aim to also establish a recommended resolution for rubric grade boundaries. Students (n = 51) carried out five separate week-long design projects. They then undertook self and peer assessments and were assessed by the module tutors, by additional staff (both technical and academic not linked to the project module) and students on other years of the course. All staff had reasonable agreement with tutor grades, although there were mixed results on whether these were a valid means to accurately assess the project outcome-depending on which statistical analysis was adopted. For all staff, >90% of their assessments were within 10% of the benchmark (agreed) grades, while for student self-Assessments 84.2% were within 10% and for student peer assessments 60.5% of the overall grades were within 10% of the tutors grades. When we analysed the data for individual tutor marks for all assessment criteria, 37.3% of grades were within 5% of each other, 85.8% were within 10% of each other, and the range within which 95% of grades could be bound was 20%, suggesting that the resolution of our rubric grade boundaries should be increased to 20% between grades boundaries for individual criteria in such practical assessment tasks in design. More generally, understanding the reliability and validity of assessment processes can help inform the assessment design to ensure an appropriate resolution is used for criteria.