In a case-control study of alcohol consumption and oesophageal cancer, investigators match cases and controls on tobacco smoking because it is correlated with drinking. Later analysis shows the odds ratio is much closer to unity than in prior studies. The most likely explanation is:
- A Neyman's bias has inflated the true effect
- B Overmatching has biased the estimate toward the null ✓
- C The confidence interval was too narrow
- D Reverse causation has masked the association
Correct answer: B. Overmatching has biased the estimate toward the null
Explanation
Overmatching occurs when a matching variable is itself associated with the exposure of interest rather than being a pure confounder. Matching on such a variable removes real differences in exposure between cases and controls, driving the odds ratio toward the null and reducing statistical power. Neyman's bias concerns survival-related selection of prevalent cases and reverse causation refers to outcome influencing exposure, neither of which applies here.
Reference: Gordis Epidemiology, 6th ed.
High-yield for: NEET PGINI-CETNExTFMGEUSMLEPLABMRCP
Written and medically reviewed by the StethoPrep medical team.