Experimento em parcelas subdivididas com tratamentos secundários em blocos incompletos balanceados

• Autor(a) principal: Vilas Bôas, Isolina Aparecida
• Data de Publicação: 2018
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFLA
• Texto Completo: http://repositorio.ufla.br/jspui/handle/1/34704
• topic: Análise de variância; Planejamento de experimento; Delineamento experimental; Sistema de equação normal (SEN); Análise intrabloco; Analysis of variance; Experimental planning; Experimental design; Sistem of normal equation; Intrablock analysis; Teoria Geral e Processos Estocásticos

This study aimed to develop a methodology for analysis of variance of experiments in subdivided plots with secondary treatments arranged according to a structure of incomplete balanced blocks, as well as to propose a linear model and to develop the procedures to perform the analysis of variance, present the expressions for comparisons of and illustrate the application of the methodology by means of an example of cultivar of corn. In the development of the methodology, it was assumed an experiment with split plots in which u secondary treatments were present in s the incomplete balanced b blocks (s < b) where each block had k plots and each pair of secondary treatments occurred in l blocks. Using the least squares method, the system of normal equations, solutions to the effects of major treatments, the effects adjusted for secondary treatment and interaction and the criteria for multiple comparisons by the Tukey test based on the variances of the linear functions estimated. Checked that two situations were characterized: the effects of the primary treatments are determined by similar to those found in the literature for complete randomized blocks, and effects related to secondary treatments and interaction were similar to those found for incomplete block experiments. As an illustration of the methodology was applied to a set of data, which were analyzed by three models and using the statistical programs R and SAS.