How to run a study group that actually helps you learn

A study group works when everyone arrives having already worked through the material and the meeting is spent quizzing each other, explaining mistakes and sorting out questions. Three to five people, an agenda and one goal per session are enough. Reading notes aloud as a group uses up the time without testing anyone.

The advantage of a group is that someone always asks the question you would not have asked yourself. Roediger and Karpicke (2006) showed with prose passages that answering questions about what you studied led to better recall a week later than studying it again. A well-run group turns that practice into a routine with a fixed time.

With two people a session easily turns into a chat; with more than five, each person speaks little and scheduling gets hard. Look for classmates in the same course or preparing for the same exam, with a similar level of commitment. They do not need to be at your level: explaining to someone slightly behind also forces you to organize what you know.

Agree on frequency, length and place from the start, in person or on a video call. A fixed weekly session, such as Tuesdays from 6:00 to 7:00 p.m., is easier to keep than hunting for a slot every week. Also decide what happens when someone shows up unprepared.

Group work amplifies individual preparation; it does not replace it. If nobody has read the topic, the meeting becomes an improvised lecture. Give each member a small task that can be checked.

This agenda is our suggestion for a one-hour session. The timings are approximate: what matters is that most of the time goes to answering without notes and correcting.

Roles stop one person from ending up as the permanent teacher. One person moderates and keeps time, another writes down the questions the group cannot answer, and a third prepares the hard problem. Rotate the roles so everyone practices explaining, not just listening.

Dunlosky and colleagues (2013) rated practice testing as high utility and rereading as low utility. Study groups were not one of the techniques in that review, but the conclusion still shapes the design: spend the time on questions and explanations, not on going through the same slides together.

When someone misses a question, ask them how they reached their answer before giving the correct one. That way the group finds the exact step that broke: a confused concept, a forgotten condition or a calculation. Handing over the solution is faster, but it leaves the mistake in place.

If two people defend different answers, open the source instead of voting. If the source does not settle it, the question goes on the instructor list. Our own example: in statistics, someone says the median of 2, 4, 6 and 100 is 28; the group checks that 28 is the mean, and the median is 5, the average of 4 and 6.

Some groups get comfortable and stop being useful without anyone saying so. Check once a month whether the sessions help you answer better, and change the format before giving up on the group.

In Lumia AI, books can be private, public or shared; whether you can share a book or make it public depends on your plan's permissions, so check the pricing page before building the group around that option. Even with a shared book, each person rates their own reviews Again, Hard, Good or Easy based on what they recalled.

For the question round, you can generate a custom quiz from a book, choosing the number of questions, the difficulty and types such as multiple choice or true or false. Review the generated questions before using them and treat the score as one more practice round: it does not certify that the group is ready for the exam.