A nonparametric test of stationarity for independent data

A nonparametric test of stationarity for independent data is investigated. The test is based on comparing kernel density estimates calculated from subsamples of the data. Asymptotic distribution theory is developed and results of a modest simulation study are presented.