Quizzing each other works, and it fails in two predictable ways

The tool below builds the set: paste your questions and it hands them back one at a time with the answers hidden, ready to read out. Nothing is uploaded and nobody needs an account to play from your screen.

Why being asked aloud beats reading together

Two friends reading the same chapter side by side are doing two separate reading sessions in one room. Two friends asking each other questions are each producing answers from memory, which is a different cognitive event with a different result: the testing effect is one of the most consistently replicated findings in the area, and the classic demonstration by Roediger and Karpicke found the advantage over rereading grew rather than shrank as the delay to the final test increased. The audience adds something on top. Being watched raises effort on tasks you already know how to do, an old result called social facilitation, and the same presence makes it harder to skip a question by telling yourself you would have got it.

Failure one, the group that answers as a group

When three people are asked a question together, they produce fewer correct answers between them than the same three would have produced separately, a robust laboratory finding usually explained by one person's answer disrupting everyone else's retrieval path before it completes. The practical version is that whoever speaks first sets the direction and the other two stop searching. The fix is procedural rather than motivational: ask the question, make everyone write an answer down before anyone speaks, then compare. This is exactly the structure of peer instruction, where students commit to an answer, argue with a neighbour, then commit again, and the second vote is the one that improves.

Failure two, the person who stops trying

In any group task, individual effort tends to fall as group size rises, the effect known as social loafing, first measured by Ringelmann with people pulling on a rope. In a study group it looks like agreement: the quiet member nods along, gets asked nothing directly, and leaves having retrieved nothing at all. Identifiable individual contribution is what removes it, so rotate who answers rather than letting the room answer, and keep a visible record of who has been asked what. Groups of four and above almost always need a turn order to survive this. Broader group dynamics and the structure of team based learning both make the same recommendation.

Writing a set that survives being played

A question with an obvious giveaway is not a question. Three rules cover most of it. Keep the options plausible, since an item where three of four choices are absurd tests reading rather than knowledge, a standard concern in multiple choice item writing. Avoid questions whose answer is another question's stem, which happens constantly when a set is written in one sitting from one chapter. And write at least a third of the set as open questions with no options at all, because recognising the right option and producing the answer are different abilities and only one of them is what the exam will ask for. The tool above keeps options attached to their question precisely so an item stays intact when you read it out.

Scoring that does not simply reward the fastest reader

Speed scoring is the default in most quiz apps and it quietly changes what is being practised: whoever reads fastest wins, and the person who thinks before answering is punished for it. Two fixes work. Score the round on correctness alone and use time only to break ties. Or set a fixed answering window, long enough that reading speed is not the constraint, and take every answer at the end of it. If you want the competition to stay close across sessions, rate players rather than counting wins, which is what the Elo rating system does in chess: beating someone far below you moves your number barely at all, so a strong player gains nothing by picking easy opponents. Anything built on points per second will drift toward gamification that rewards the game rather than the learning.

A classroom is the same problem at a different scale

Everything above holds for a room of thirty, with one addition: the teacher needs to see the distribution of answers, not just the winner. That is what audience response systems were built for, and what makes a quiz round a piece of formative assessment rather than a game. The useful signal is not the score but which wrong option attracted a third of the class, because that is a misconception with a name and an address. Shared sets also spread the writing cost: collaborative learning works best when each person authors a section and is answerable for it. There is more on classroom use on the teachers page.

Duelling on your own material in the app

Upload a PDF, a photograph of a page or a camera shot and WeSolve+ writes an explained quiz from that single file, then lets you duel a friend on those questions rather than on a generic bank. Wrong answers collect in their own list, so a duel you lose tells you exactly what to work on. Sets can be shared into a classroom. It runs at wesolveapp.com in any browser and as an App Store app on iPhone and iPad. Free to start, with two quizzes and two uploads a day, two recaps a week and two duels a day. There is no separate Android app. Numbers on pricing, and the mechanics on duel.

Last updated: 2026-08-15