The computational approach tests for testing the equality of population means under two-parameter exponential distribution

Testing equality of the several populations' mean is one of the main statistical problems. Classical F test is the most powerful test for the solution of the problem when the assumptions are normality and variance homogeneity. These assumpti- ons may be violated in practice. In this case, the researchers improved alternative solutions using normality transformation techniques and robust estimators. However, these solutions may not give powerful results in all cases. The problem has a large set of possible solutions when the normality assumption is violated. There are some tests have been improved using Generalized p-Value, Parametric Bootstrap and Fiducial Approach methods for log-normal, inverse-normal and two-parameter exponential distribution in the presence of nuisance parameter in the literature. In this study, new tests based on Computational Approach method are proposed to solve the problem of testing the equality of two-parameter exponentially distributed populations' means. The performance of the proposed tests is compared with the alternatives in the literature. As a result, the proposed tests performed better es- pecially in small samples. In addition, the proposed tests have been applied on real data sets, and thus their advantages over the alternatives have been demonstrated.