Checklist
How to avoid sampling bias: seven checks
Seven checks for keeping sampling bias out of a study: define the population and sampling frame, select at random where possible, use quotas or stratified sampling, recruit across channels, include churned and inactive users, follow up with non-responders and size the sample so each subgroup can be weighted. Use it when planning recruitment for interviews or surveys. With Koji you can run many interviews at once, so filling every segment does not stretch the timeline.
Where this appears
What's in this checklist
- Define your target population and match your sampling frame to it
- Use random selection wherever you can
- Apply quotas or stratified sampling so no group is crowded out
- Recruit through several channels and go after hard-to-reach groups
- Interview churned and inactive users, not only active ones
- Follow up with a subset of non-responders to see how they differ
- Grow the sample so every subgroup can be represented and weighted
Use this image
Free to use in your docs, decks and posts under CC BY 4.0, as long as you credit Koji with a link to this page. To embed it, paste this into your page:
html
<a href="https://www.koji.so/visuals/how-to-avoid-sampling-bias-checklist"><img src="https://www.koji.so/visuals/how-to-avoid-sampling-bias-checklist/how-to-avoid-sampling-bias-checklist.png" alt="Checklist of seven ways to avoid sampling bias, from defining the sampling frame and random selection to quotas, churned users and non-responders." width="1080" height="1080" /></a>
<p>Source: <a href="https://www.koji.so/visuals/how-to-avoid-sampling-bias-checklist">How to avoid sampling bias: seven checks</a>, Koji</p>
