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 hospital records should be examined?
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