Intervalos de confianza Jackknife para cuantiles en muestreo con probabilidades desiguales

Abstract

We consider the estimation of quantiles in finite population susing the jackknife technique. We use a jackknife variance estimator for unequal probability sampling that works better than the classical jackknife estimator. The quality of the confidence interval found is demonstrated via simulation. The proposed confidence interval showed coverage probabilities that were close to the nominal confidence level, and mean lengths and variances much smaller than the mean lengths and variances of the intervals using the traditional jackknife methodology.