Abstract
Tests and confidence intervals for the error variance in a standard regression model are obtained, using weighted jackkife method. The resulting inferences are shown to be robust against nonnormality. A weighted jackknife version of Spjøtvoll's (1967) test of against
in the unbalanced random one-way model
is developed (where {vi} and {eij} are mutually independent random variables with mean zero and variances
respectively. Power comparisons are made in a simulation study with the unweighted jackknife version of Arvesen and Layard (1975).