A pollster asks people “coffee or tea?”. Each answer counts 100 (coffee) or 0 (tea), so one answer wobbles by about spread points around the town’s true share (a standard deviation, from “How spread out?”). A poll’s result is the average over the n people asked. Write wobble_of_average(spread, n): how much does a poll’s result wobble from one poll to the next?
A town where really half prefer coffee and half tea. A pollster asks 25 people at random, again and again, and reports the share saying coffee.
wobble_of_average(spread, n)| spread | how much one person's answer wobbles, in points (a standard deviation) | e.g. 50 |
| n | how many people the poll averages | e.g. 1 |
| returns | how much the poll's result wobbles | |