Years ago now when I was working at the SU at UEA, someone gave me access to that year’s National Student Survey (NSS) free-text comments.
As my eyes skimmed the text, one of the things that stuck out like a sore thumb was the number of students specifically mentioning “group work” and/or “group assessment”.
And not in a good way.
A few years later on eve of the NSS’ 15th birthday, the Office for Students (OfS) promised that it would undertake detailed analysis of trends, looking at some of the key themes emerging from the open-ended question.
It’s one of the many useful things that OfS has promised to do over the years but has seemingly never got around to.
I raise all of this because the other day, something happened down under that has drawn derision from inside the sector.
Bye Bye Bye
Speaking at the Universities Australia annual shindig, shadow education spokesperson Julian Leeser grabbed the headlines by arguing that universities should remove group assignments altogether.
He argued students “hate them” and that they are “deeply unfair”, framed them as undermining standards in way that “cheapens the degree”, and linked it to a wider critique of assessment integrity and individual learning.
He also raised concerns about concentrations of international students, questioned whether degrees still reflect “intellectual effort” and suggested more invigilated assessment and on-campus attendance.
Here he is, speaking to the yoof on insta.
Education minister Jason Clare had rocked up the same event, and in the press conference a journalist asked him directly whether group assignments should be scrapped. He replied with a political joke first, saying the Liberal Party seemed to have trouble with teamwork, and then gave the substantive answer that universities should decide – but that teamwork is a real skill because most jobs require people to “work as part of a team” and “you can achieve more as a team than you can as an individual.”
Much of the rest of the commentary has tended towards the usual autonomy/expertise/ridicule thing – but Leeser isn’t daft. The student who said to me on Wednesday that they’d had to submit a “group essay” – and that it had been the worst experience of his university life – had a point.
2 Become 1
We’re not sat on the data mountain that OfS is, but we have been running student experience polling for a few years now, with plenty of opportunities to communicate the “why” behind the Likert ticks.
So I had a proper look.
Nearly two-thirds of the respondents who answered an assessment format question in one of the waves reported having used group work or collaborative outputs as a form of assessed work. It’s a widely experienced format.
Among those who rated their confidence that group work allows them to show genuine learning, opinion was sharply divided – just over 30 per cent were fairly or very confident, but 52 per cent were not very or not at all confident.
Where we’ve asked about assessment, group work is the least trusted assessment format in the survey by a clear margin – over a third of students who have used it lack confidence that it shows genuine learning, and it sits bottom of all twelve formats rated.
The core complaint is fairness, not pedagogy – students resent shared marks when contributions are unequal, and the frustration tracks with wider dissatisfaction with marking fairness, feedback quality, and a perceived gap between what courses say they value and what they actually reward.
The pattern is concentrated among undergraduates in their middle and final years, particularly in Business and Creative Arts, and is associated with notably worse happiness and higher anxiety.
No Scrubs
Then across the open-ended questions, hundreds of comments speak directly to group work as an assessment issue.
The dominant complaint was fairness and free-riding. Students describe one or two people doing all the work while others contribute nothing, then everyone receiving the same mark.
Comments like “one person doing the work for everyone,” “some people get marks they don’t deserve for others’ hard work,” and “group mates didn’t contribute much or not at all” recur across multiple questions.
Several frame this explicitly in terms of assessment validity – the mark doesn’t reflect what they individually know or can do.
Another cluster argues that group work doesn’t show individual learning, and many argue that even when a group functions reasonably well, the format doesn’t accurately represent each person’s ability:
“Not a fair portrayal of how one individual has learnt – impossible to tell who did how much.
Grading mechanisms are part of the problem. Plenty draw a careful distinction between group learning (which they value) and group grading (which they resent). Proposals include grading individual contributions, having tutors observe the process rather than just the output, requiring peer review that actually affects marks, and attributing sections to individual members.
One student explicitly said that group work “clearly improves group skills” but shouldn’t count towards final grades.
Group presentations were a particular flashpoint. Comments about group presentations come up frequently and negatively – students describe them as especially difficult to navigate, unfair in grading, and want them replaced with either individual presentations or different group formats.
Say My Name
In The Conversation, Jason Lodge (who is Director of the Learning, Instruction & Technology Lab at the University of Queensland) says the quiet part out loud – that group assignments lessen the marking and feedback load, particularly in courses with high numbers of students:
“For cash-strapped universities, the efficiency is hard to ignore.
He goes on to argue that nevertheless, group assignments can serve both pragmatic purposes (reducing marking load) and pedagogic ones – particularly in disciplines like health where interprofessional teamwork is an accreditation requirement, not an optional extra.
He draws on his own lab’s research showing students spend as much time negotiating how to do the work as doing it, and that this “co-regulated learning” – involving emotional regulation, problem-solving and planning – is itself a crucial skill that becomes more rather than less important in an age of AI.
His conclusion is that the debate should be about improving group assessment design to recognise the process as well as the product, not about abolishing it.
Wider research backs him up. Michael Thom’s 2020 review in Teaching Public Administration found that most studies on group assignments suffer from weak research designs, leaving little empirical basis for their presumed benefits, and that students consistently report not enjoying or valuing them.
He argues that universities should stop treating group work as a default, coordinate across programmes to prevent students being hit with group projects in every module simultaneously, and should invest properly in teaching collaborative skills rather than assuming students already have them.
Meanwhile, meta-analyses drawing on larger evidence bases – Springer, Stanne and Donovan in 1999 across STEM, Johnson, Johnson and Smith in 2014 more broadly – do show gains in achievement, attitudes and persistence from structured cooperative learning. But the critical word is “structured.”
What the meta-analyses measured was not the kind of unstructured group essay that students in our polling are complaining about.
The research on what makes group work actually function points to a set of design features that are well understood but widely ignored. The task needs to be complex enough that members genuinely need each other – not something that one competent person could do alone in a weekend.
Students need to be explicitly taught how to work in teams, manage conflict and communicate professionally, rather than being dropped into groups and left to figure it out. Groups that have autonomy over how they organise their work report stronger collaboration.
And shorter, more focused collaborative tasks appear to produce better outcomes than the long, sprawling group projects that generate the most complaints. A study across multiple courses at Utrecht University found that the density and complexity of the task, combined with its perceived relevance to students’ professional development, were the strongest predictors of effective collaboration. Groups producing outputs they saw as meaningful – a genuine research proposal, a report for an external client – engaged more deeply than those completing work they perceived as artificial.
On the fairness issue specifically, the sector has tools it is largely failing to use. Peer assessment of individual contribution – where group members rate each other and the ratings generate a weighting factor that modifies the group mark – is well-established in the research and supported by dedicated platforms.
WebPA, developed at Loughborough, and SPARKPlus from Australia both allow criteria-based peer moderation of group marks at scale, with SPARKPlus specifically designed to detect free-riders and non-contributors.
UCL’s IPAC Consortium, spanning over 20 departments, has found that 92 per cent of students support peer assessment of contribution when it is implemented properly – with the key condition being that feedback is shared transparently rather than used as a hidden moderation tool. Work from the design education field has found that peer weighting left over 90 per cent of students satisfied that their grade reflected their actual performance.
A 2025 study went further, proposing that peer ratings be combined with digital trace data from collaborative platforms to create composite measures of individual input – addressing the concern that peer assessment alone can be gamed or distorted by social dynamics.
The point is not that these approaches are perfect, but that they exist, they work, and the students in our data who are asking for individual contribution to be recognised are asking for something the literature has already solved.
Harder Better Faster Stronger
Back in Australia, QAA/OfS hybrid TEQSA (Tertiary Education Quality and Standards Agency) has argued that in an age of AI, universities now need more sophisticated ways of assuring learning. It means more of an emphasis on things only humans can do – not less.
That includes collaboration. The paper explicitly argues that assessment should create “opportunities for good quality collaborative work”, with clear rules on what counts as acceptable collaboration with both peers and AI, and with marks informed by evidence of the process as well as the product.
In other words, the problem the students are describing – opaque contributions, shared marks, unclear boundaries – is not an argument against group work, but against poorly specified group work. Where collaboration is assessed, it needs to be designed, observed and evidenced.
Seen that way, the politics looks a bit thin. Scrapping group assignments might remove a source of student frustration, but it would also remove one of the few places in a degree where students are asked to demonstrate how they work with others under pressure.
Ignoring deep dissatisfaction from students while sneering at an opposition intervention is one option. The other would be to treat group work not as a convenient way to reduce marking workload, but as a capability to be taught, structured and assessed properly.
Thanx for this.
Moots are a form of group work that seem to be greatly enjoyed by law students. The essential feature is that each member of the group – senior counsel, junior counsel, and solicitor – has a well defined role and students are evaluated on how well they perform their role. Indeed, one of the points of moots is to develop students’ understanding of the different roles of counsel, solicitor, and judge.
I find myself in the middle ground here. Group projects can be excellent; this is often a sign of good group cohesion. Most presentations, however, are a bit ‘what I did in my holidays’, with little or no connection between different parts. The great advantage group presentations have over individual presentations is that it means that students who, for whatever reasons, do not want to speak in public can still take part.
I like the idea of a collaborative discussion of something not fully discussed in the class, but I don’t find the results the most useful in terms of awarding a final grade for a course.
Nice AI slop image
That’s interesting—and it’s worth noting that assessors don’t necessarily favour these mechanisms either, for a range of reasons.
In group work, for example, familiar problems persist: free-riders are often visible (particularly to other students), but so too are more subtle dynamics—such as domineering individuals who take over and marginalise others, often without recognising their own behaviour. Isolation within groups can be a real issue, and in some cases this can extend into forms of exclusion or even bullying.
Peer review mechanisms, while well-intentioned, can sometimes reinforce rather than resolve these dynamics. They rely on students accurately and fairly evaluating one another, which is not always realistic given these interpersonal complexities.
As one student once put it to me: “I came to university to be marked by a professor, not my classmates.” That sentiment captures an important concern about where authority and responsibility for judgement should sit.
In my advanced seminar, students have to present (a part of) a research paper that they’ve identified and read that’s relevant to both their final report for the class and of interest to the broader student membership of the class. They are explicitly allowed to work individually or in groups of up to three — the bigger the group, the longer they have to present and the higher the expectation of the depth/difficulty of what they present. It is entirely up to the students to self-organize into groups if they want to work together.
I have had zero complaints about this sort of group work, and often a lot of expressions of gratitude for it, because it allows those who want to work together to do something bigger to do so.