Randomness in the Classroom: A Primer
Where random selection helps a lesson, where it hurts one, and what teachers report.
Random selection in a classroom does two useful things and one harmful one, and the balance depends heavily on how it is used rather than on whether it is used.
The useful things are well established: it distributes participation away from the same few hands, and it removes the teacher from the selection, which changes how being called on is received.
The harmful thing is that being called on unpredictably is genuinely stressful for some students, and a tool that increases participation by increasing anxiety has not straightforwardly improved anything.
What random calling actually changes
Without a mechanism, participation concentrates. The same students volunteer, the teacher's attention follows familiarity, and a substantial fraction of a class can go a full term contributing nothing verbally. That is not a controversial observation and it is the problem random calling addresses.
The second effect is on the students being called. A selection nobody made is received differently from one the teacher made — it removes the reading that being asked means the teacher thinks you were not listening, which is a common and corrosive interpretation.
The third effect is on preparation. If any student might be asked, the incentive to have thought about the question applies to everyone rather than to the volunteers, which is the mechanism most of the pedagogical argument rests on.
The anxiety problem, honestly
For some students, unpredictable public questioning is a significant source of stress rather than a mild one, and the effect is largest for exactly the students who most avoid volunteering. A technique that increases participation by making a quiet student anxious has traded one problem for another.
There are well-described mitigations and they cost very little. Thinking time before any name is drawn, allowing a pass with a return later, permitting a student to confer with a neighbour first, and warning that a question is coming all reduce the stress substantially while keeping the participation benefit.
The mitigation most teachers find most effective is the simplest: draw the name after the question and after the thinking time, not before. Everyone prepares an answer, and only then does the selection happen.
- Ask the question before drawing the name.
- Give explicit thinking time to everyone before the draw.
- Allow a pass with an explicit return later in the lesson.
- Allow conferring with a neighbour before answering.
Coverage versus equal chance
There is a real design decision here and it is not obvious. An independent draw gives every student an equal chance every time; a no-repeat cycle guarantees everyone is asked once before anyone is asked twice.
Independent draws leave gaps. Over thirty draws from a class of thirty, an expected eleven students are never picked — more than a third of the class, while somebody is picked four times. That is not a defect of the draw, it is what independence means, and it defeats the entire purpose of using a mechanism.
A cycle guarantees coverage and gives up equal odds: the last draw of a cycle is a certainty rather than a one-in-thirty. For a classroom that trade is almost always correct, because coverage was the goal and equal odds never were.
Thirty draws from a class of thirty
- Independent draws: an expected 11 students never picked; someone picked 3 or 4 times.
- No-repeat cycle: every student picked exactly once.
- The cycle gives up equal odds within the cycle to buy that guarantee.
The rigged-picker complaint
Every teacher who uses a picker eventually gets told it is rigged, usually by a student who has been picked twice in a week. The complaint is predictable and the arithmetic is reassuring.
In a class of thirty over sixty lessons with independent draws, a specific student being picked twice in a row has probability 1 in 900 per adjacent pair — but across all students and all adjacent pairs, at least one such double is close to certain. The class sees the class of coincidences; the student sees their specific one.
The two responses that work are computing that number in front of them, and switching to a visible cycle where the shrinking pool is on screen. The second works better with younger classes, because it replaces an argument with an observation.
Where randomness is the wrong tool
Seating is the clearest case. A fully random plan ignores sight lines, hearing, support needs and the pairs who must not sit together, all of which are non-negotiable — so the randomised portion should be whatever is left after those are fixed, not the whole room.
Assessment is another. Randomising which students are assessed, or which questions each gets, undermines comparability, and comparability is usually the point. Randomising the order in which work is marked is useful and different.
And group formation for sustained project work is a case where random allocation is defensible and often worse than deliberate allocation, because the cost of a badly composed group is borne over weeks rather than minutes.
Frequently asked questions
Does random calling actually increase participation?
It distributes it. Without a mechanism, participation concentrates on the same few students and a substantial fraction of a class can contribute nothing verbally over a term.
Is it stressful for students?
For some, significantly — and largest for the students who most avoid volunteering. That cost is real and is why the mitigations matter.
What is the single most effective mitigation?
Ask the question and give thinking time before drawing the name. Everyone prepares an answer, and only then does selection happen.
Should I use an independent draw or a no-repeat cycle?
A cycle, almost always. Thirty independent draws from thirty students leave an expected eleven never picked, which defeats the purpose.
What is the cost of a cycle?
Unequal odds within it. The first draw is one in thirty and the last is a certainty — that is what buys the coverage guarantee.
How do I answer 'the picker is rigged'?
Compute how often the run they saw should occur, or switch to a visible cycle where the shrinking pool is on screen. The second works better with younger classes.
Should seating be fully random?
No. Sight lines, hearing and support needs are non-negotiable, so fix those seats first and randomise what remains.
Is random group allocation better than choosing?
For short activities, usually. For sustained project work the cost of a badly composed group is borne over weeks, and deliberate allocation is often better.
Tools that use this
Related guides
Why Humans Are Bad at Being Random
Asked to be random, people avoid repeats, alternate too often, and dodge the edges.
Bracket and Fixture Scheduling Maths
Byes, seeding, and the circle method that fits a full round-robin into n-1 rounds.
Constraint Satisfaction in Fair Draws
What happens when the rules and the randomness disagree — and when there is no answer.