The Use of the Jackknife to Estimate Proportions from Toxicological Data in the Presence of Litter Effects

The jackknifing methodology is proposed as a way of estimating and comparing proportions of affected fetuses in animal experiments. The estimate is a weighted average of the proportions in individual litters and is almost fully efficient in several cases of interest. The associated standard error is asymptotically correct for a wide range of possible models. Jackknife-based tests are shown to be competitive with other methods currently used.