Abstract
SIR—Floated variances should be used routinely in epidemiology. The way most epidemiologists now calculate confidence intervals for relative risks is not appropriate, and, as a result, the intervals are often much too wide. This difficulty affects a very large number of epidemiological reports. It arises whenever a comparison is being made where one group is designated as the reference category (therefore having a relative risk of 1), and relative risks are calculated that compare each of the other groups with this one. The standard confidence intervals for such relative risks should generally be replaced by confidence intervals that are based on the floated variances of the log odds ratios for the groups that are being compared with each other. 1,2
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
