A fair random student picker is not a lottery you run fresh every lesson — it is a rotation. Put every student's name in once, pick from it without putting names back, and only reshuffle after the whole class has been called. The math behind the plan is simple: sessions needed = class size ÷ picks per session, rounded up. A class of 30 students with 4 picks per session needs 8 sessions (7 sessions of 4, plus a final session of 2), and at the end of the cycle every student has been called exactly once. That is the difference between random and fair: randomness decides the order, the rotation guarantees the coverage.
Pure random picking, where every name goes back in the hat after each draw, feels fair but is not. With replacement, repeats are normal and omissions are common. In a class of 30, if you draw 30 names with replacement, the chance that any one particular student is never drawn is (29/30)^30 — about 36%. Across a whole class, that means roughly 11 of 30 students would never be called at all in those 30 draws, while a few unlucky students get called three or four times. Students notice that pattern fast, and once they decide the system is rigged, participation drops. The fix costs nothing: draw without replacement until the list is empty, then start a new cycle.
Here is a worked example. A teacher has 30 students and wants 4 cold calls per lesson, with 2 minutes per answer in a 10-minute slot. Sessions needed = 30 ÷ 4 = 7.5, rounded up to 8. Sessions 1 through 7 call 4 students each (28 students), and session 8 calls the remaining 2. If the class meets 5 days a week, the full cycle takes just under two weeks. Measurely's Random Student Picker does exactly this arithmetic: enter class size and picks per session, and it returns the session count, the final session size, and the rotation plan. You still draw the actual names from a hat, cards, or a shuffled list each session — the tool plans the fairness, the draw supplies the surprise.
Choosing picks per session is a time decision, not a fairness decision. Work backward from the minutes you can spare: divide the discussion minutes by the minutes per answer. A 10-minute slot at 2 minutes per answer gives 5 picks; a 15-minute slot at 3 minutes per answer gives 5 as well. Fewer picks per session stretches the cycle across more lessons, which is fine — a longer cycle is still fair, as long as you finish it before reshuffling. What breaks fairness is abandoning a cycle halfway because you forgot where you were, so keep the list visible: cross names off on the board or in a spreadsheet.
Absences are the one routine exception, and they have a simple rule: skip the absent student and slot them into the next session, not the end of the cycle. If a student misses their turn in session 3, they become the first pick of session 4, and the rotation continues from there. This keeps the cycle length unchanged and avoids the quiet unfairness of an absent student drifting to the back of every queue. The same rule covers a student who is present but genuinely unprepared for a question that requires preparation: move them one session, not to the bottom, and be open about the rule so the class sees it applied the same way to everyone.
Random calling works best when it is paired with the right kind of question. Quick recall and opinion prompts suit cold calls well, because any student can answer something. Questions that need preparation — a passage to read, a problem to solve first — should be announced in advance, or given as think-pair-share before the draw. And when the job is group work rather than individual answers, switch tools: the Random Group Generator splits the same class into even groups, which is a different fairness problem (equal group sizes, not equal turns) and deserves its own plan rather than a stretched picker rotation.
Three mistakes undo an otherwise fair system. First, reshuffling mid-cycle because the remaining names feel boring — that resets coverage and recreates the repeat problem. Second, re-drawing a name you already called because the first student gave a short answer — one call is one call, whatever the answer. Third, letting volunteers jump the queue: volunteering is great, but a volunteer answer should not consume a rotation slot, or the quiet students subsidize the confident ones. Write the two rules on the board at the start of term — no repeats until everyone has been called, and volunteers answer in addition to the draw, not instead of it.
A final checklist before your first cycle. Count the class and set picks per session from your available minutes. Build the name list once and draw without replacement. Keep the crossed-off list where you can see it. Slot absent students into the next session. Finish the cycle before reshuffling. Run your numbers through the Random Student Picker — 30 students and 4 picks per session is 8 sessions, a cycle you can finish in under two weeks — and random calling stops being a complaint and starts being the fairest routine in the room.