Recall bias
- Bias
People who are ill tend to search for and remember their past exposures more than healthy people do, and that inflates the association.
What it is
Recall bias is a differential information bias: participants remember past exposures with different accuracy depending on their outcome. It is typical of case-control studies, where cases looking for an explanation of their illness remember or report more exposures than controls. It usually exaggerates the odds ratio and is prevented by using records instead of memory.
How it arises
- Illness makes a person search their past for causes.
- Healthy controls have no such reason and forget more exposures.
- Exposure ends up under-recorded in controls or over-recorded in cases.
Worked example
- Study of 150 mothers of babies with a malformation (cases) and 150 of healthy babies (controls); exposure: fever in pregnancy.
- By questionnaire: fever reported by 45 of 150 cases and 24 of 150 controls.
- Questionnaire OR = (45 × 126) / (105 × 24) = 5 670 / 2 520 = 2.25.
- From antenatal records: fever in 30 of 150 cases and 24 of 150 controls.
- Records OR = (30 × 126) / (120 × 24) = 3 780 / 2 880 ≈ 1.3.
- Reading: mothers of cases recalled more episodes; memory inflated the OR from 1.3 to 2.25.
Markers
- OR
- Differential
How to prevent or control it
- Whenever possible, take exposure from records made before diagnosis.
- Use structured questionnaires, the same for cases and controls, and interviewers who do not know who is a case.
- A study nested in a cohort with data collected at baseline avoids this bias.
Common pitfalls
- Thinking recall bias is simply forgetting: what biases is that the error differs between cases and controls.
- Blaming it on prospective cohorts: there, exposure is measured before the outcome.
Test yourself
Think of the answer, then tap the question to see it.
In which design is recall bias most typical?
In case-control studies, when exposure is obtained by asking participants.
Which way does recall bias usually move the odds ratio?
Away from 1: it usually exaggerates the association, because cases report more exposures than controls.
How is recall bias prevented?
By using records or samples from before diagnosis, the same structured questionnaires for everyone and blinded interviewers.
Connections
Educational content on epidemiological and statistical methods; the examples use made-up numbers to teach the calculation. Not a clinical or legal guide: current Colombian rules are in the official sources.
Sources: Recall bias | Catalog of Bias (Catalogue of Bias Collaboration, 2017); Principles of Epidemiology | Lesson 1 - Section 7 (Centers for Disease Control and Prevention, 2012).
Spotted a mistake?We are always improving for the community. Tell us what is wrong and our team will review it.Report a mistake
Representative model, for educational purposes only.