Odds ratio, relative risk and case-control studies

Bob Wheeler

bwheelerg@gmail.com

Many studies in the social sciences make use of the odds ratio to summarize data in 2x2 tables. In most cases, the odds ratio is used as an approximation to a statistic of greater interest, the relative risk, which is not always calculable because of the method of sampling. The justification for this is that they are close when the underlying risks are small, which is true; but since in the usual course of research this is not verifiable and since the difference in the risks is actually more important, odds ratios are reported that are sometimes quite different than the relative risk. In any case, contrary to the widespread belief, relative risk may be calculated in many studies without the need to use the odds ratio as an approximation.

Full File in pdf format.