Traditional methods for testing equality of means are based on normality observations in each treatment, but parametric bootstrap methods offer a test statistic to estimate P-value by resampling. In article, first, Fisher, Cochran, Welch, James, Brown and Forsyth, Approximate F, Weerahandi, Adjust Welch and Parametric Bootstrap tests for testing hypothesis equality of means are defined. Then type one error and power of these tests were compared to each other by a simulation study for various sizes of samples and treatments. Finally sizes of these tests were calculated for the real data of Esfahan Cement factory.
Traditional methods for testing equality of means are based on normality observations in each treatment, but parametric bootstrap methods offer a test statistic to estimate P-value by resampling. In article, first, Fisher, Cochran, Welch, James, Brown and Forsyth, Approximate F, Weerahandi, Adjust Welch and Parametric Bootstrap tests for testing hypothesis equality of means are defined. Then type one error and power of these tests were compared to each other by a simulation study for various sizes of samples and treatments. Finally sizes of these tests were calculated for the real data of Esfahan Cement factory.

Keywords: Analysis of variance, Behrens-Fisher problem, Parametric bootstrap, Monte Carlo simulation

Full-Text [PDF 908 kb]   (1875 Downloads)