I am about to ask MTV viewers the following question: “Do you regularly watch MTV’s Beavis and Butthead?” Last week I asked 100 of my friends the same question and 35 of them said yes.
How many actual viewers must be surveyed to be 95% confident that the estimate of the viewership is in error by at most 3%? (Assume 100% response rate.)
What would be your answer if I had not surveyed any of my friends?
This solution gives the step by step method for computing sample size for estimating population proportion.