Selection bias
Etymology
From "selection" (Latin "seligere," to choose out) and "bias" (Old French "biais," a slant). It became a standard term in 20th-century statistics and epidemiology for errors introduced by non-random sampling.
noun phrase · /suh-LEK-shun BYE-us/
A distortion that arises when the sample you study isn’t representative of the whole.
What it names
Explains why data can mislead — if who or what gets included is skewed, the conclusion is skewed before you begin.
Example
A survey about internet habits taken only online suffers from selection bias — it can’t hear from people who aren’t there.
Related idea
Survivorship bias is a vivid special case: studying only the winners or survivors, blind to those who dropped out and left no trace.