Cold-calling is the teaching move of asking a student who did not volunteer, so the whole class stays ready to think instead of a few hands carrying every question. It has two familiar failure modes: it can feel like a targeted gotcha, and it can quietly become unfair when the teacher keeps landing on the same students. A random student selector answers both at once — the choice is visibly out of your hands, and it is spread evenly across the room.
The bias a selector removes
When you pick a name in the half-second after a question, you are not picking evenly. Attention slides to the students with their hands up, the ones near the front, and the names that surface fastest — while the quiet middle of the room can go a whole lesson without being asked. None of that is deliberate, which is exactly why trying harder does not fix it. A selector treats every present student as one equal entry, so being called no longer depends on being eager or memorable.
How to use it for a cold call
Build your class list once and keep it as a roster you reopen each lesson. Turn on no-repeat for the round so every present student is called exactly once before anyone comes up again — that answers “why always me” and “never me” at the same time. Mark absent students as not present so they are skipped for the session without losing their place. Then change the order most teachers use by habit: ask the question first, pause, and only then let the selector say a name — so every student prepares an answer, not just the one who gets picked.
Keep the cold call warm
A selector decides who is picked, not what happens next — and cold-calling earns trust only when it stays supportive. Give real wait time once a name appears, keep your tone matter-of-fact rather than a gotcha, and leave an honest way through: let a student think aloud, pass and come back, or bring in a partner. Used that way the randomness lowers the stakes, because being chosen is plainly the luck of the draw and not a spotlight you aimed.
Is the selector actually even?
“Random” is free to print and impossible to check by eye, so we measured it. Running WheelMigo's real selection function 7,000,000 times over a seven-name list, every name landed within 0.015 of a percentage point of a perfect one-in-seven — because it draws from the browser's crypto.getRandomValues and uses rejection sampling to strip the modulo bias that skews naive pickers. For a single spin that gap is invisible either way; for a class you cold-call every day, an even selector is the difference between fair over time and a standing advantage no one can see on screen.