By Cyril KimbiMost study groups fail, and they fail in a recognisable way: four people sit together, read their own notes, talk about the exam, and go home having done what they would have done alone but with more interruptions. A group that works looks nothing like that, and the difference is entirely structural.
What a group is actually for
Two things, and only two.
Explaining exposes gaps that silent revision hides. When you read a topic you understand it fluently, which feels like knowing it. When you explain it to someone who does not know it, the exact sentence where your explanation goes vague is the exact point you do not understand. Nothing else finds that boundary so quickly.
Someone else marks you honestly. Marking your own work, you know what you meant. A group member reading your answer against the marking scheme only knows what you wrote — which is all the examiner will know either.
Everything else a group does — company, motivation, sharing notes — is a side benefit, not a reason. If a session delivers neither explaining nor marking, it was a social occasion.
The size and the composition
Three or four people. Two is a conversation and drifts. Six is an audience, and in an audience most people stop participating.
Mixed ability works better than matched ability, in both directions. The stronger student consolidates by teaching — explaining is among the most effective ways to secure your own knowledge. The weaker student gets an explanation pitched by someone who learned it recently, which is often clearer than the teacher’s.
What does not work is a group where one person always explains and the others always listen. That is a class, and the listeners are passive. Rotate.
A session format that works
Ninety minutes, one topic, decided in advance.
- 10 minutes — blank-page recall. Everyone writes what they remember about the topic, individually, no notes. Silent.
- 25 minutes — teaching round. Each person explains one part from memory. The others interrupt with questions whenever an explanation goes vague. Being interrupted is the point.
- 30 minutes — past-paper questions, individually and timed. Same questions, silence, clock running.
- 20 minutes — cross-marking. Swap papers. Mark someone else’s against the marking scheme, strictly. Explain each mark you withheld.
- 5 minutes — name the next topic and who prepares which part.
The cross-marking step is where the value concentrates, and it is the step groups drop first because it is uncomfortable. Marking a friend strictly feels unkind. Marking them generously is what actually costs them.
The rules that keep it working
- One topic per session, named in advance. “We will do physics” produces a session about physics. “Projectile motion, 2015-2019 questions” produces work.
- Come prepared or do not come. A group that carries unprepared members becomes a class, then becomes nothing.
- Phones in a pile, face down, away from the table. One phone answered restarts three people’s concentration.
- Fixed day and time. A group that reschedules weekly dissolves within a month.
- Start on time regardless of who is missing. Waiting punishes the punctual and teaches lateness.
- Ninety minutes, then stop. Sessions that run long are usually sessions that started late and drifted.
Groups without reliable connectivity
This is where a study group has an advantage over solo work rather than a disadvantage, and it is worth using deliberately.
Split the downloading. Four people each fetch a different subject’s papers and marking schemes, then share the files offline. Each has spent a quarter of the data and holds the whole library. That alone can be worth more than the sessions.
The method needs nothing. Explaining aloud, timed papers and cross-marking require paper, a pen and a clock. A group meeting under a tree with printed papers is doing the highest-value activity available to any candidate anywhere.
See the offline study guide for how to structure the fetching, and the past questions library for the material itself.
When a group is the wrong answer
Be honest about this. A group is not useful when you are learning a topic for the first time — initial learning is individual, and a group meeting about material nobody has studied is four people being confused together. It is also the wrong tool if the group’s real function is reassurance, which is comfortable and produces nothing.
The sequence that works is: learn alone, consolidate together. Study the topic individually first, then bring it to the group to be tested on it.
Frequently asked questions
How big should a study group be?
Three or four. Two drifts into conversation; six turns most members into an audience.
How often should we meet?
Once or twice a week, same day and time. Frequency matters less than reliability.
What if one person contributes nothing?
Address it directly and early. A passenger changes the group’s norms faster than anyone expects.
Should we all study the same subject?
For a session, yes. A session where everyone works on something different is four people doing solo work in one room.
Is a group better than studying alone?
Neither is better. They do different jobs: alone for learning, together for explaining and being marked.
