Randy Neil Biostats 08: USMLE Step 1 - Bias & Confounding with Questions

USMLE purpose: Distinguish the major epidemiology biases and recognize when a third variable is a confounder.

How to use this page

Start with the bias recognition table and memorize the classic clue for each bias.
Use the confounding rule: the third variable must affect both exposure and outcome.
Test yourself with the rapid recall and practice questions before opening the answer toggles.
Use the transcript only if you want Randy’s original wording.

Bias questions are pattern-recognition questions: identify the study flaw, then match the clue to the named bias.

Bias / concept	Classic clue	Board-style takeaway
Recall bias	Questionnaire after the outcome occurred	Cases may remember exposures differently than controls
Selection bias	Nonrepresentative sample	Chosen group differs systematically from target population
Hawthorne effect	Participants know they are observed	Behavior changes because of observation
Pygmalion effect	High expectations change performance	Expectation can alter outcomes
Procedural bias	Pressure affects survey/evaluation response	Respondents do not answer freely
Lead-time bias	Earlier detection	Survival appears longer without delaying death
Length-time bias	Slow-growing disease	Screening overdetects indolent disease
Confounding	Third variable affects exposure and outcome	Control with stratification, matching, restriction, regression, or randomization

Questionnaire after matching/nonmatching outcome → recall bias.
Slowly progressive tumor → length-time bias.
Early detection makes survival look longer → lead-time bias.
Third variable related to both exposure and outcome → confounding.
Crossover study → patient acts as their own control; washout period matters.

⚠️ Trap: Calling every third variable a confounder.

Fix: A true confounder must be associated with both the exposure and the outcome.