Member Training: The Fundamentals of Sample Size Calculations

Sample size estimates are one of those data analysis tasks that look straightforward, but once you try to do one, make you want to bang your head against the computer in frustration. Or, maybe that’s just me.

Regardless of how they make you feel, they are super important to do for your study before you collect the data.

As tempting as it is to skip or fudge this task, it will only lead to bigger headaches later on when data collection takes forever or you discover you don’t have enough data to test your hypotheses.

So how do you actually do them, and how to do you understand the many, many issues involved?

In this month’s Stat’s Amore webinar, we will lay out the seven steps to run an accurate sample size estimate. Some of them are super easy (like #2: choose an alpha level). Others are much harder to do in practice than they look (#3: specify the smallest effect size of scientific interest).

We will also discuss some fundamental concepts that you need to really wrap your head around. They will help you understand how to approach these seven steps. And more importantly they’ll help you understand why you must avoid some of the common but bad practices that are often recommended by well-meaning advisers and what to do when you’ve hit a tricky sample-size situation.