Adjustment of an Odds Ratio for Selection Bias


Denote the observed data from a case-control study in the form of a traditional epidemiological 2x2 table using: a,b,c,d. Denote the analogous population contingencies as: A,B,C,D. Here, there are four selection probabilities resulting from the four disease-exposure contingencies: a/A, b/B, c/C, d/D. A quick way to evaluate the impact of selection bias in a case-control study is to calculate a "selection odds ratio": [(a/A)*(d/D)] / [(b/B)*(c/C)]. Multiplying an observed odds ratio by the inverse of this selection odds ratio will result in an adjusted estimate.

Explore the Impact of Selection Bias by Putting Some Values into this Calculator

Observed Odds Ratio:

Selection Probability (diseased, exposed):

Selection Probability (non diseased, exposed):

Selection Probability (diseased, non-exposed):

Selection Probability (non diseased, non-exposed):