Skip to content
BENIBIX Learn

Selection Bias

General🔬 ResearchResearch Quality

Also known as Sampling Bias, Selection Effect

When the people picked for a study do not match the group it speaks for.

Think of it like

It is like asking kids in a pizza shop if they like pizza. Of course they do.

Example

A website asks its readers if they shop online. Almost all say yes. The readers are already online, so the answer says little about everyone.

Why it matters

A study can be done with great care and still mislead. The wrong people were in it.

Definition

Selection bias happens when the people in a study are not like the group it claims to describe. It comes from how people were picked or who chose to join. Then the result is true for the sample but not for everyone. It can make a treatment look better or worse than it is. Random picking is the main defence against it.
Left panel: only the people near the door get picked. So the sample is lopsided.

Common mistake

Thinking a big sample fixes it. A million people picked the wrong way are still the wrong people.

Key takeaways

  1. 1
    Who got in. Picking shapes the result.
  2. 2
    Self-chosen groups. Volunteers differ from others.
  3. 3
    Pick at random. The main defence.