Questions: Matching in Case-Control Studies

5 questions to test your understanding

Score: 0 / 5
Question 1 Multiple Choice

A researcher conducts a matched case-control study, pairing each case with a control of the same age and sex. She then analyzes the data using standard (unmatched) logistic regression. What is the main problem with this approach?

AStandard logistic regression cannot handle binary outcomes
BThe analysis ignores the pairing structure, treating matched controls as if randomly selected and producing biased estimates
CThe matched variables (age, sex) will not appear in the output, so confounding is uncontrolled
DStandard logistic regression overfits when sample sizes are small
Question 2 Multiple Choice

A researcher studying smoking and bladder cancer matches each case to a control on age, sex, and whether the person has nicotine-stained fingers. What is the likely consequence of matching on nicotine-stained fingers?

AStatistical power will increase because a strong confounder has been controlled
BConfounding by nicotine will be eliminated, improving validity
COvermatching will occur — cases and controls will be similarly exposed, reducing the detectable exposure contrast
DThe matched analysis will require a larger matched set ratio to compensate
Question 3 True / False

Matching on a confounding variable in a case-control study eliminates the need for statistical adjustment of that variable in the analysis.

TTrue
FFalse
Question 4 True / False

Matching in case-control studies is generally preferable to statistical adjustment because it controls confounding more mostly.

TTrue
FFalse
Question 5 Short Answer

Why does matched data require a matched analysis, and what happens statistically if you ignore the matching?

Think about your answer, then reveal below.