Consider k independent random samples with different sample sizes such that the ith sample comes from the cumulative distribution function (cdf) Fi = Gαi, where αi is a known positive constant and G is an absolutely continuous cdf. This paper shows how one can construct the nonparametric confidence intervals for the quantiles of G. Toward this end, we consider two cases. First we assume that the complete samples are available and then suppose that only the maxima and minima have been observed. In each case exact expressions for the confidence coefficients are derived. Two mentioned procedures are compared via numerical computations and simulation study.

Keywords