How to Run a Study Group That Is Actually Worth the Time

A familiar pattern: five people agree to meet, everyone is enthusiastic, and forty minutes later it is a conversation about something else entirely.

People conclude from this that group study does not work. It works, but only for particular things, and the failure is almost always structural rather than about discipline.

A group is excellent at finding out what you do not know. It is poor at fixing it. Once you sort activities by that rule, everything else about running a group becomes obvious.

What groups are genuinely good at

Do together	Do alone
Explaining a topic aloud to each other	Reviewing cards on a schedule
Questioning each other from past papers	Learning new material for the first time
Comparing how you each answered a question	Timed full papers
Arguing about which method is correct	Reading and note making
Finding out what you cannot explain	Fixing what you could not explain

The left column all share one property: they require another person. The right column does not, and doing it in a group means five people working slowly in parallel while distracting each other.

A group that struggles is often doing right column activities together.

The one rule that fixes most of it

Decide the topic before the session, not at it.

A group that arrives without a specific topic spends the first twenty minutes deciding, which is the hardest and least rewarding part, done at the point everyone is least focused. Then it drifts, because nothing was agreed to drift away from.

Agree the topic at the end of the previous session, or the day before in a message. One topic, named specifically. Not chemistry, but reactions of alkenes. Everyone arrives having read it, which turns the session into practice rather than teaching.

Rotate who explains

The single most valuable activity available to a group, and the one people avoid because it is uncomfortable.

One person explains the topic to the others, from memory, without notes, for five minutes. Everyone else asks questions until something breaks. Then swap.

Two things happen. The explainer discovers within seconds what they cannot actually articulate, which silent reading hides completely. And the listeners discover which questions they did not know to ask.

Rotate every session. A group where the same confident person always explains is a group where one person is learning and four are watching.

Keep it small

Aim for three or four people. That number is a rule of thumb, not a research finding: the UNC guide only says to keep the group small enough for everyone to ask and answer questions. In a large group, quiet members tend to disappear.

The people should be sitting the same subjects, roughly matched in level. A group where one person is much stronger turns into a lesson, which is useful for four of them and a waste of an evening for the fifth. If that is the arrangement, be honest about it and treat it as tutoring.

Use a marking scheme, not opinions

The most productive thing a group can do with a past paper is mark each other’s answers against an actual scheme.

You will be far stricter with someone else’s work than with your own, which is exactly the point. And after struggling to find someone’s third point buried in a paragraph, everyone starts putting each point on its own line without needing to be told.

How to read a scheme properly, and how to mark strictly enough for it to be worth doing, is in our guide to how marking schemes work.

Set an end time

Ninety minutes, then stop. A session with no end drifts into conversation and everyone leaves feeling they wasted an afternoon, which makes the next one less likely to happen.

The UNC Learning Center suggests agreeing the length in advance, at 60 to 90 minutes, and meeting about once a week.

Finish by agreeing the next topic and the next time. Thirty seconds, and it is the difference between a group and a one off meeting.

Turn what you found into your own revision

This step is easy to skip, and skipping it is how a group can feel productive without changing anyone’s results.

The group told you what you cannot explain. That information is worthless unless it becomes something you review. Before you leave, write down every gap the session exposed. Then turn each one into a card and review it on a schedule, using the method in our twelve week revision plan.

Diagnosis in the group, treatment on your own.

A session that actually works

Ninety minutes, one topic, three or four people. This shape survives contact with reality better than anything more ambitious.

Minutes	What happens
0 to 10	Everyone writes down what they remember of the topic, alone, on a blank page
10 to 30	One person explains it from memory. The others question until something breaks
30 to 55	Past paper questions on the topic, attempted alone, timed
55 to 80	Swap papers and mark each other against the scheme
80 to 90	Each person writes their own gaps. Agree next topic and time. Stop

A good deal of that is silent, individual work done in the same room. That is deliberate. The group provides the structure and the accountability, and the actual thinking still happens in one head at a time.

The last ten minutes are not optional. A session that ends without each person writing their own gaps has produced a pleasant afternoon and no change to anyone’s revision.

What the research says about each part of the 90 minutes

The session plan above was designed to put the group’s strengths to work. This table checks each step against the sources checked on 18 October 2026 and says how strong the support is. Most of the evidence is about individual techniques; very little tests a self-run revision group.

Step	What the evidence says	Strength
0 to 10: blank-page recall	Practice testing is one of two techniques Dunlosky and colleagues rated high utility (Psychological Science in the Public Interest, 2013). A meta-analysis of the testing effect by Rowland (2014) found that initial recall tests gave larger benefits than recognition tests.	Strong, for individuals
10 to 30: explain from memory	In Koh and colleagues’ 2018 experiment, as summarised in a 2022 review, students who explained a text on video without referring to it, or who simply did a free-recall test, outscored students who read a script aloud (d = 0.57 and 0.65) and students who did a filler task (d = 0.82 and 0.91). Explaining and plain recall did not differ much (d = 0.10). The review concludes the evidence is still limited and the retrieval-versus-explanation question is open.	Moderate. Supports doing it from memory; does not prove an audience adds anything
30 to 55: timed questions, alone	Same practice-testing evidence as above. The NHS also suggests practising past papers under exam conditions (NHS exam stress).	Strong
55 to 80: swap and mark	No study testing peer marking against a mark scheme was found in the sources checked. The UNC guide lists creating likely exam questions and answering them together among useful group activities.	Untested here. A reasonable practice, not a proven one
80 to 90: gaps become cards	Distributed (spaced) practice is the other technique Dunlosky and colleagues rated high utility.	Strong, for individuals
Size, length, frequency	The UNC Learning Center says groups work best when small enough for everyone to ask and answer questions, suggests about one meeting a week, and suggests 60 to 90 minutes. It does not give a number of members. The 3 or 4 in this post is a rule of thumb.	Guidance, not measurement

There is also broader support for small-group learning. A 1999 meta-analysis by Springer, Stanne and Donovan of 39 studies (undergraduate science, mathematics, engineering and technology courses since 1980) concluded that various forms of small-group learning promote achievement, better attitudes and persistence. Treat that with care: these were studies of university science, maths, engineering and technology courses, not of students organising themselves for a school-leaving exam.

What you can take from all of this is a modest conclusion. The individual steps rest on solid evidence, the group-specific steps rest on sensible guidance, and nothing here promises a particular grade. If you try the plan, write down in each session how many gaps you found; if that number is not falling after a month, change something.

The failure modes, and what causes them

What goes wrong	The actual cause
Drifts into conversation	No topic agreed in advance
One person teaches, others listen	Nobody rotates the explaining role
Feels productive, results unchanged	Gaps never written down or reviewed after
Attendance falls away	Sessions too long or too frequent
Quiet members contribute nothing	Group too large
Arguments about who is right	No marking scheme to settle it

Every row has a structural fix rather than requiring anyone to try harder.

If there is nobody around you

Plenty of candidates are the only person in their circle sitting a particular subject, or are preparing outside a school entirely.

Online groups fill that gap. Many candidates are working through the same syllabus at the same time, and finding them costs nothing. Our community page covers the eStudyUniverse GCE channel, which is broadcast only and low volume, and the Advanced Level study group, where candidates ask each other questions.

The rules above still apply online, and one more matters more there: mute the group while you work and check it deliberately afterwards. A chat open beside you during revision is a distraction wearing the clothes of study.

Staying safe in an online group

Worth stating plainly, because many members of student groups are minors.

Nobody in a genuine study group needs your money. Anyone selling leaked papers, guaranteed results or admission is defrauding you, and leaked papers are almost always fake as well as a serious offence. Never send money, bank details or identity documents to someone you met in a group, and if anyone messages you privately in a way that makes you uncomfortable, leave and tell an adult you trust.

Also verify anything specific. Dates, cut off marks and requirements circulate in student groups constantly and are frequently out of date or invented. Confirm at the source before acting on any of it, for the reasons set out in why exam information circulating online is so often wrong.

Questions students ask

How often should we meet?

Once a week is enough and sustainable. More than that and preparation becomes a burden, attendance drops, and the group quietly ends. A weekly group that runs for six months beats a daily one that lasts three weeks.

Someone never prepares. What do we do?

Address it directly and early rather than resenting it. Rotating the explaining role usually solves it on its own, because arriving unprepared becomes visible. If it continues, the group is better without them, and saying so is kinder than the group dissolving instead.

Is it better than studying alone?

Not better, different. Most of your revision should be alone, because that is where memory is built. A group is a weekly diagnostic that tells you where to point the alone time.

Can we study together over WhatsApp instead of meeting?

Yes, and voice notes work well for the explaining exercise, since the point is producing an explanation aloud rather than being in a room. Agree a fixed time so it is a session rather than a chat that never quite happens.

We are in different towns. Can it still work?

Yes, and it is common now. A voice call for the explaining exercise, attempting the same past paper separately at an agreed time, then photographing your answers and marking each other’s works well and needs almost no data if you agree the paper in advance and download it once.

Set the first one up this week

Find two or three people sitting your subjects. Agree one topic, one time, ninety minutes. Whoever knows the topic least explains it first.

If you cannot find anyone locally, start with the online community, where other candidates are preparing for the same papers you are.

More method in Study Skills.