Two-sample comparisons for serially correlated data / M.B. Seitshiro

The purpose of this study is to derive new tests for the equality of the means in twoindependent or dependent stationary time series, based on bootstrap critical values.Required properties of these tests include satisfactory probability of Type I errors, andhigh power. It is shown how critical points for various sample sizes and significancelevels can be obtained by applying the parametric bootstrap. A limited Monte Carlosimulation study is conducted to illustrate the validity of the bootstrap approximationof the exact critical values, by producing satisfactory probability of Type I errors. Italso shows that the newly proposed tests compare favourably with standard two-sampletests in the absence of serial correlation, under the null hypothesis of equalmeans, but are more powerful than the well-known t -test if small and moderatecorrelation structures are present, for a wide range of parameter values. All findingsand conclusions of the Monte Carlo simulations are reported.