EFFICIENCY OF MEAN SEPARATION TESTS IN OAT TRIALS
Abstract
The use of mean separation tests is one of the alternatives to compare treatments, when the factors are qualitative and not structured. This paper had the objective to evaluate the use of mean separation tests in oat trials and to discuss the advantages of application of more appropriate statistical procedures. The data analyzed were published in the annals of "Reunião da Comissão Brasileira de Pesquisa de Aveia" (Meeting of Oat Research Brazilian Committee), from 1996 to 1998. The published papers were classified as follows: appropriate, partially appropriate and inappropriate, concerning its use. The conclusions regarding the oat trials were encouraging. More than 80% of the papers that presented mean separation tests were properly applied, with a tendency to increase in the evaluated years. The mean separation tests applied inadequately represented 10% (in average) of analyzed papers, where the most frequent error was to apply these tests to compare levels of quantitative factors. The incorrect use of mean separation tests does not imply, necessarily, in erroneous conclusions. However, it may limit the inferences that could be made on the obtained data.
Downloads
The authors declare that the work has not been previously published, nor sent simultaneously for publication in another journal and that they agree with the submission, content and transfer of the publication rights of the article in question to the scientific journal Pesquisa Agropecuária Gaúcha - PAG. The authors assume full responsibility for the originality of the article, and may incur on them any charges arising from claims by third parties in relation to the authorship of the article. The full reproduction of the journal's articles in other free-to-use electronic media is permitted under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license.
