Hypothesis testing for factorial effects in replicated orthogonal experiment
This paper studies the testing method for data from orthogonal experiments with heteroge-neous variances.In the data analysis of orthogonal experiments,the F test is one of the most used methods to test the significance of the effects of the factors.In the Ftest,it is usually assumed that the variances of the errors under different treatments are equal to each other.However,it is not always so in practice.This paper proposes a new testing method for the occasions that the variances of the errors under different treat-ments are unequal to each other.Simulation studies show that,under such occasions,the new method can control the probability of type one error more accurately than the F test.
国家科技期刊平台
NSTL
万方数据
