Grades in formative workplace-based assessment: a study of what works for whom and why

Grades are commonly used in formative workplace-based assessment (WBA) in medical education and training, but may draw attention away from feedback about the task. A dilemma arises because the self-regulatory focus of a trainee must include self-awareness relative to agreed standards, which implies grading.

In this study we aimed to understand the meaning which medical students construct from WBA feedback with and without grades, and what influences this.

Year 3 students were invited to take part in a randomised crossover study in which each student served as his or her own control. Each student undertook one WBA with and one without grades, and then chose whether or not to be given grades in a third WBA. These preferences were explored in semi-structured interviews. A realist approach to analysis was used to gain understanding of student preferences and the impact of feedback with and without grades.

Of 83 students who were given feedback with and without grades, 65 (78%) then chose to have feedback with grades and 18 (22%) without grades in their third WBA. A total of 24 students were interviewed. Students described how grades locate their performance and calibrate their self-assessment. For some, low grades focused attention and effort. Satisfactory and high grades enhanced self-efficacy.

Grades are concrete, powerful and blunt, can be harmful and need to be explained to help students create helpful meaning from them. Low grades risk reducing self-efficacy in some and may encourage others to focus on proving their ability rather than on improvement. A metaphor of the semi-permeable membrane is introduced to elucidate how students reduced potential negative effects and enhanced the positive effects of feedback with grades by selective filtering and pumping. This study illuminates the complexity of the processing of feedback by its recipients, and informs the use of grading in the provision of more effective, tailored feedback.