Jackknife estimators of variance for parameter estimates from estimating equations with applications to clustered survival data.

摘要: An estimate of a parameter vector beta is often obtained by setting "score" equal to zero and solving for beta. Estimating equations this type include maximum likelihood, quasi-likelihood (McCullagh, 1983, Annals Statistics 11, 59-67), generalized estimating (Liang Zeger, 1986, Biometrika 73, 13-22). White (1982, Econometrica 50, 1-26) proposed variance estimator that robust model misspecification. We show "one-step" jackknife asymptotically equivalent the White. The one-step may be preferred when appropriate computer packages are not available compute White's directly. This very useful in our example with clustered survival data.