On the generalized jackknife and its relation to statistical differentials

Abstract

SUMMARY The flexibility of the definition of the first-order generalized jackknife is exploited so that its relation to the method of statistical differentials can be seen. The estimators presented have the same bias reduction and asymptotic distributional properties as the usual generalized jackknife. A limiting case produces an infinitesimal jackknife which represents a generalization of statistical differentials.