The one mode that works: being made to explain
Explaining a concept to someone who does not understand it is among the most demanding things you can do with knowledge, because they interrupt at exactly the point you were going to gloss over. You cannot use the textbook's phrasing as cover — they'll ask what it means. It is the Feynman technique with an enforcement mechanism that talks back.
The person doing the explaining learns far more than the person listening. Which has a consequence people find unwelcome: in a study group, you want to be the one who doesn't understand it yet, and then to become the one explaining it. Groups where the same person always explains are groups where one person is studying and the rest are attending.
The modes that don't work
- Parallel silent working. This is co-working. It's fine for accountability and does nothing for learning. Be honest that that's what it is.
- Dividing the syllabus. "You do chapters 1–3, I'll do 4–6, we'll swap summaries." Five people each learn one fifth and receive four summaries they will not retain. It feels like the efficient move and it is the single worst thing a study group can do.
- Group re-reading. Reading the same slides together, out loud. Now four people are experiencing the fluency illusion simultaneously.
How to run one that actually works
- Everyone studies the topic alone, first. The group is not where you learn it; it's where you test it.
- Take turns explaining, without notes. Blank whiteboard. The rest of the group's job is to ask 'why?' relentlessly.
- Quiz each other on things you didn't prepare. Prepared questions are the ones you already know.
- Pool the material, never the learning. Share a folder so everyone works from the same sources — then each of you does all the retrieval practice, alone.