This is the authors' abstract. We don't add key points for this paper.
In plant breeding trials the mean phenotypic responses of two experimental genotypes are often close; and which genotype appears to perform better depends on the specific environments that are observed. When only one of the genotypes must be selected, the conclusion drawn by comparing means across any set of observed environments may differ from the conclusion that would be drawn if all target environments could have been observed, in which case the wrong selection may be made. This paper proposes a new method of comparing genotypes that aims to select the genotype that is more likely to perform better across a set of environments, rather than the one that has the better mean. The implementation uses bootstrap resampling to estimate the probability that one genotype outperforms another in a set of observed environments, and by doing so accounts for the uncertainty caused by observing limited environments. The results show that due to the different genetic-by-environment (G×E) interaction effects, the genotype that is more likely to be better is sometimes different than the one with the better mean and the probabilistic comparison accounts for both the mean and the interaction effects.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Bijari et al. (2022) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: