Suppose a study estimates the relative risk of disease for those persons with a binary condition compared to persons without the condition to be 3.1, with a 95% CI of 2.4 to 5.3. Why is this confidence interval not symmetric around the sample relative risk of 3.1? (ie: the endpoints not equal distance from the endpoints)

a. This is a mistake: the 95% CI for a relative risk must be symmetric around the sample estimate.

b. Because this msut have been based on a small sample, and hence a t-correction was used to create the confidence interval.

c. Confidence intervals for ratios are not necessarily symmettic around the sample estimate on the ratio scale, but they are on ln(ratio) scale.

1point