Okay here is my two cents on sample size. Suppose you have research questions and you need to know how many participants to recruit for your study. You want to ensure that your study has enough ‘evidence’ to be viewed as having enough power. When I refer to accuracy or ‘power of the test’ I am talking about whether or not your results correctly reject/fail to reject your null hypothesis. In general, a large sample should give you more ‘power of the test’ (a lot of studies aim for at least 80% power). Once you know how many participants will be in your study, you can proceed from that point. It’s a great way to ensure you have the best representation of your results. All is good and groovy. You can take your results, and ride into the sunset.
But you should have an idea of what statistical method you will perform BEFORE you try to calculate your sample size. Any statistics consultant can assist you formulating your hypothesis and statistical methods. You should not be confused about the statistical method as you calculate your sample size.
-If you enjoyed this blog there are more to follow.-Amy