How much data do I need to collect?

In the previous page, we investigated how to determine the sample size needed to estimate a population mean to a specified accuracy. A similar calculation can be used to find the size of sample required for estimating a probability.

A small pilot survey is often conducted to obtain a preliminary estimate for the proportion.

If we can do no better, the 'worst-case' value, p = 0.5 can be used, but the resulting sample size may be higher than needed.

How many people should be phoned?

This example finds the sample size needed to estimate a proportion with specified accuracy.

Obtaining the sample size by solving an equation

Trial-and-error can be avoided with a little algebra. The equation

can be re-written in the form