Imagine you are sampling a population, and you want to know how many you need to ask in order to have a reasonable confidence that the answer you get is credible for the whole population. This tool calculates the sample you will need, depending on the underlying population, and your expected response rate, and the level of confidence you want around it.

The uncertainty is a combination of the acceptable range (the error) and the confidence you want that the sample is representative. But it is also affected by the size of the underlying population, and the expected response rate.

Don’t worry if you don’t know the exact population – once it gets quite big, it doesn’t matter so much you can just estimate. But it’s important to think between