In a case-control study of coffee drinking and pancreatic cancer, investigators match controls to cases on sex, age, and also on non-drinking of alcohol. The association between coffee and pancreatic cancer becomes much harder to detect. The most likely reason is:
- A Recall bias introduced by matching
- B Loss of statistical power due to unequal group sizes
- C Overmatching on a variable in the causal chain ✓
- D Confounding created by the matching variable
Correct answer: C. Overmatching on a variable in the causal chain
Explanation
Matching on a factor that lies on the causal pathway between exposure and disease, or that is strongly associated with the exposure itself, is overmatching. Because coffee drinkers and abstainers differ systematically in alcohol habits, forcing equality on alcohol removes real exposure contrast and drives the odds ratio toward the null. Matching never causes confounding; its purpose is to control it, so option D states the opposite of what occurs.
Reference: Gordis Epidemiology, 6th ed.
High-yield for: NEET PGINI-CETNExTFMGEUSMLEPLABMRCP
Written and medically reviewed by the StethoPrep medical team.