A note on the bias and consistency of the jackknife variance estimator in stratified samples