An Investigation of Robustness of Rank Tests in AMMI Multiplicative Terms

In this paper, Monte Carlo simulation technique was performed to determine the Type I error rates and power for the Additive Main effects and Multiplicative Interaction (AMMI) model multiplicative term tests for both original and rank observations (i.e., replaced the original observations with their respective ranks), and then the robust tests among them identified. The Results indicate that the Goodman and Hrbarman tests (FGH1 and FGH2) on ranks are robust for both validity and efficiency with more than 2×2 genotypes by environments combinations, while FGH1 and FGH2 tests on the original observations are non-robust. The results also indicate that Gollob-test is robust for both validity and efficiency for only 2×2 genotypes by environments combination. The Type I error rates of the Gollob-test, generally increases with the increase in the number of genotype by environment combinations.