Introduction

You'll be disappointed if you jumped to this Handbook to get a one number answer to the question “what is the perfect sample size?”. There isn't a perfect sample size. Determining an appropriate sample size can be fickle and contingent on many factors. Many factors (like someone’s willingness to participate) can be out of your hands. Unless you continue to recruit/contact participants to reach a certain number, your study’s sample size is one of the least controlled aspects of any study you run.

But there’s a problem: you can’t guarantee the desired sample size without spending resources, and your stakeholders generally trust your findings when the sample size is large. Yes, larger sample sizes generally increase confidence in your research. But size is a brute force characteristic, meaning that adding more and more participants doesn't guarantee higher research quality or more impactful learning. As you’ll read in Topic 2 and Topic 3, large samples can sometimes be horrible and small samples can sometimes be more than enough.

How you determine, explain, and defend the sample size you set out to study is what matters. This Handbook does contain some suggestions for qualitative and quantitative research sample sizes. But what this Handbook can’t do is recognize if these suggestions can actually work in your context. You’ll have to understand the reasoning behind each suggestion for them to make sense. Once they do, you can explain and debunk hesitations or expectations and align on practical sample size. A practical sample size, as defined here, is just enough for your stakeholders to see the value in the research and for you to study sustainably.

One quick note: The terms small and large are used heavily in this Handbook. However, small and large are relative terms; small for a startup might look wildly different than small for a big tech company. Use your best judgment to interpret what small and large mean in your context.

Before addressing why large sample sizes aren’t always great, let’s start at the other end: why small samples aren’t inherently bad.