A ratio of these quantities can be directly computed, while relative risk can only be indirectly inferred from a case-control study. For 10 points each:
[10m] Name this measure of exposure–outcome association that, for a 2-by-2 contingency table, equals a times d over b times c. The log of this ratio is the dependent variable of logistic regression.
ANSWER: odds [or odds ratio; accept OR; prompt on p/(1 − p)]
[10e] Odds ratios are subject to the selection form of this phenomenon, since the selected sample does not accurately represent the population of interest.
ANSWER: bias [accept selection bias]
[10h] For a case-cohort study, relative risk approximately equals the odds ratio under this epidemiological assumption first identified by Jerome Cornfield. This assumption is unnecessary when performing “incidence-density” sampling.
ANSWER: the rare disease assumption [accept answers referring to the prevalence or incidence or probability of having the disease or outcome or condition being small; accept answers specifying that both a and c are small or negligible]
<GC, Other Science>