Journal of Educational Assessment Science and Collaborative Learning Critique
Reeves, D., Park, M., Oyelaran, S.
Department of Educational Assessment Science, University of Ashford
Applied Pedagogy Research Unit, Harwick University
Received: 05 March 2025 · Accepted: 05 March 2025
This study examines the relationship between group project assessment structures and individual performance accuracy in educational settings. The Group Assessment Equity Index (GAEI) was administered to 314 students across individual and group assessment conditions. Results indicate that group project grades accurately reflect individual contribution in 31% of cases. In the remaining 69%, the grade either over-rewards low contributors or under-rewards high contributors, producing what the authors term 'contribution noise' — a grading outcome decoupled from individual performance. The group project's stated aim is collaborative learning. Its documented outcome is collaborative grade-sharing, which is a different thing entirely.
Group projects are a standard pedagogical tool across educational levels, premised on the dual objective of developing collaborative skills and producing outcomes that reflect collective intellectual effort (Johnson & Johnson, 2009). Their assessment structure — typically a single grade applied uniformly to all group members — assumes a level of contribution equity that the social loafing literature suggests is structurally unlikely in voluntary collaborative contexts (Latané, Williams & Harkins, 1979). The present study provides the first application of the GAEI to group project assessment outcomes, treating the group grade not as a measure of individual learning but as an organizational output whose relationship to individual performance is highly variable and measurable. We examine who benefits, who is harmed, and whether the educational institution issuing the grade has any mechanism for knowing which is which.
Participants.
Three hundred and fourteen undergraduate students (M age = 20.3, SD = 1.8) across 78 project groups were recruited during a standard semester. Exclusion criteria included projects with fewer than three members and self-assembled groups where pre-existing friendship structures might have resolved the coordination problem before it arose (n = 7 groups, excluded as insufficiently realistic). IRB protocol EA-2024-0149 was approved.
Instruments.
The Group Assessment Equity Index (GAEI; 23 items, α = .92) measured individual contribution, perceived equity of grade distribution, effort investment, and what students described as 'watching the person do nothing and calculating their GPA implication.' Contribution was assessed via multi-source peer rating, output tracking, and meeting attendance logs. A control group completed equivalent individual assessments and received grades reflecting, uniquely, what they had done.
Procedure.
GAEI was administered at project midpoint and completion.
Grade-Contribution Accuracy.
Group project grades accurately reflected individual contribution levels in 31.2% of cases. In 43.1%, the grade over-rewarded low contributors. In 25.7%, it under-rewarded high contributors. Contribution variance within groups was significant (SD = 28.4 GAEI contribution points), confirming that group member performance was not homogeneous.
Social Loafing Prevalence.
Peer contribution ratings identified at least one non-contributing member (below 20th percentile contribution threshold) in 67.3% of project groups. These members received an equivalent final grade in 91.4% of cases.
High-Contributor Stress.
GAEI stress subscale scores were significantly elevated in high-contribution individuals (top quartile), driven by awareness of grade-contribution discrepancy and what 78% described as 'having to decide whether to do their part or let the grade suffer.'
The 31% grade-contribution accuracy rate establishes that group project grades are, in the majority of cases, not measuring what they claim to measure. A grade attached to a group outcome and distributed uniformly to all members is not an assessment of individual learning or effort. It is an average of the group's output, assigned irrespective of who produced it — a redistribution mechanism that is, in 69% of cases, inaccurate in one direction or the other.
The 91.4% rate at which non-contributing members receive equivalent grades confirms that the group project, as typically assessed, functions as an involuntary contribution subsidy. High contributors' effort raises the grade ceiling for those who did not produce it. This is not collaboration. It is a transfer payment with no mechanism for consent.
The 78% stress rate among high contributors — driven by the awareness of this dynamic — suggests that the educational system is not merely failing to measure accurately. It is creating a context in which the most engaged students absorb the anxiety of group performance gaps they did not create and cannot fully remedy.
Group project grades accurately measure individual contribution in 31% of cases, systematically subsidize low contributors, and produce significant stress in high contributors aware of the discrepancy. The authors recommend individual contribution tracking as a required component of all group assessments, the elimination of uniform group grading, and a frank institutional acknowledgment that 'collaborative skills' and 'doing other people's work because the grade is collective' are not the same learning objective.
Correspondence: reeves@of-ashford.ac