Métodos alternativos para estimar tamanho ótimo de parcelas experimentais: uma aplicação na cultura da bananeira

• Ano de defesa: 2011
• Autor(a) principal: Oliveira, Elisângela Aparecida de
• Orientador(a): Cecon, Paulo Roberto
• Banca de defesa: Finger, Fernando Luiz , Santos, Gérson Rodrigues dos
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Viçosa
• Programa de Pós-Graduação: Mestrado em Estatística Aplicada e Biometria
• Departamento: Estatística Aplicada e Biometria
• País: BR
• Palavras-chave em Português: Modelos segmentados; Regressão; Tamanho ótimo de parcela; Bananeira
• Palavras-chave em Inglês: Segmented model; Regression; Optimum plot size; Banana crop
• Área do conhecimento CNPq: CNPQ::CIENCIAS AGRARIAS
• Link de acesso: http://locus.ufv.br/handle/123456789/4034

The experimental unit, or plot, is the unit that receives a particular treatment and provides the data which will reflect its effect. Optimum plot size is directly related to experimental accuracy, since the appropriate size reduces the experimental error. The objectives of this study were to use the methods Linear Segmented Model with Response Plateau (MLRP), Quadratic Segmented Model with Response Plateau (MQRP), Exponential Segmented Model with Response Plateau (MERP) and the traditional Modified Maximum Curvature Method (MMCM) to determine the optimum size of experimental plots using real data of banana crop; compare the methods using the following goodness of fit indices: coefficient of determination (R2), Akaike information criterion (AIC) and Bayesian information criterion (BIC); and obtain confidence intervals for the optimum plot size. For the study, we used data from a blank experiment conducted in the municipality of Sebastiao Laranjeiras, Bahia, in two growing seasons (2006-2008), where the following traits were evaluated: plant height, pseudostem circumference, number of live leaves, number of suckers and yellow-Sigatoka score at flowering. At this evaluation, each plant was considered as a basic unit of 6 m2, totaling 240 basic units, whose combinations formed 19 different plot sizes. The results indicated that the estimates for plot sizes, in general, varied according to the evaluated characteristics, the production cycle and the used models. This variation corresponded to the range from 7 to 66 plants in the first cycle and 70 to 40 in the second cycle. The MERP model determined sizes for the basic units larger than the MQRP model, which, in turn, estimated larger sizes than the MLRP model. Based on the goodness of fit, the best model for the analyzed data was MERP, estimating a mean optimum plot size of 31 basic units. MLRP estimated mean optimum plot size of nine basic units and MQRP estimated a mean optimum plot size of 22 basic units. Statistically, the best model fitted was MERP, but for practical reasons, as itestimates larger optimum plot sizes, it may not be feasible for the researcher. The findings of this study indicate that the models MLRP, MQRP and MERP can be used to determine optimum plot size. Thus, we suggest the simultaneous use of more than one method fordetermining the optimum plot size, as it will meet the factors considered by each method and the needs of the researcher.