Edgeworth expansions of length two are established for the sample quantile, preprivoted by a smoothed bootstrap, and for a studentized sample quantile, where we studentize by means of a kernel density estimate. As a consequence, it turns out that that two-sides confidence intervals for the...