Selection bias
Say it: suh-LEK-shun BY-us
Selection bias is an error that arises when the people recruited into, or remaining in, a study differ in systematic ways from the target population.
Definition
What is selection bias?
Selection bias is a systematic error caused by the way participants are recruited, agree to take part or leave a study. If the people studied, or the groups being compared, differ from the target population in ways linked to the exposure or the outcome, the results can overstate or understate the true effect. Unlike random error, it does not shrink as the sample grows.
It takes several familiar forms. Volunteers for health research tend to be healthier and more health-conscious than people who decline, so a study of volunteers may report lower death rates than the general population. Workers usually have lower death rates than nonworkers because very ill people are less able to hold jobs, which is called the healthy worker effect. In a case–control study, choosing controls from hospital patients can distort results if their own illnesses are tied to the exposure. In cohort studies and trials, people lost to follow-up may differ from those who stay, which is called attrition bias.
Selection bias is hard to correct once data are collected, so it is mainly prevented by design: defining a clear source population, choosing cases and controls from it in the same way, randomizing with concealed allocation, keeping follow-up complete and analyzing trial participants according to the group they were randomly assigned.
Word parts
Where does the term “selection bias” come from?
- se-
- apartLatin se-
- -lect-
- to choose, gather (seligere: to choose out)Latin legere
- bias
- slant, oblique lineOld French biais
Practice it
Where you'll meet it
Quizzes that use it
Related
Terms that go with it
- ConfoundingConfounding is a distortion in a study where a third factor, linked to both the exposure and the outcome, creates, hides or exaggerates an apparent effect.
- Case–control studyA case–control study is an observational study that compares people with a disease and people without it to find past exposures linked to it.
- Cohort studyA cohort study is an observational study that follows groups differing in an exposure over time to compare how often they develop an outcome.
- Randomized controlled trialA randomized controlled trial is an experiment that assigns people by chance to a treatment or a comparison group to test whether the treatment works.
- Relative riskRelative risk is the risk of an outcome in an exposed group divided by the risk in an unexposed group, showing how strongly the two are linked.
- EpidemiologyEpidemiology is the study of how often diseases and other health events occur in populations, what causes them and how they can be controlled.
Sources
Checked against
Written for biofideX in plain words, after checking these references:
- Catalogue of Bias (Oxford CEBM): Selection bias
- NCI Dictionary of Cancer Terms: selection bias
- NCBI MeSH: Selection Bias
Last updated
Spotted a mistake? Tell us and we'll correct it.
Practice Epidemiology
Free epidemiology quizzes, with an explanation for every answer.
See the Quizzes