Each leave-one-out statistic has the same distribution as , so
Consequently the bias of the jackknife bias correction is
The assumed expansion gives
and
Their difference is therefore , as required.

Articles by others on the same topic (0)

There are currently no matching articles.