Sample size

• What is the goal?
• Is the difference expected to be large or small?
• What is an acceptable margin of error?

Transcript

Next, I need to think about sample size.  Right! First, I have to think: "What's the goal?"  And I have to think: "Is the difference that I expect to observe large or small?". Because if the difference I expected to observe is very large, I don't need very many samples to find that difference. But if the difference I'm trying to detect is very, very small, then I need a lot of samples to be able to find that.  And of course, finally - I need to think about "What is my acceptable margin of error?"  If I'm allowed to have a large margin of error, then I don't need a very large sample size. If I want to have a very small margin of error, then I'm going to need more samples.