Faixas de classificação do coeficiente de variação para avaliação da precisão em experimentos com Brachiaria spp.
Nenhuma Miniatura disponível
Data
2016-02-26
Autores
Taira, André
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal do Espírito Santo
Resumo
The coefficient of variation (CV%) is a measure of relative dispersion used traditionally for the validation experiments. Through the method proposed by Garcia (1989) sought to new CV% rating bands for the validation experiments of the most diverse species of interest and response variables in agricultural sciences, but this depended on the normal distribution of CV% 'if thus limiting the availability of classifying a given variableresponse. Subsequently to the method of Costa et al. (2002) it was possible to sort groups of CV% of unknown distributions. Due to these studies the 20% value traditionally used for the validation experiments have been replaced by other reference values in the CV%. The only CV% rating bands for existing grasses prior to this work are those of Clement and Muniz (2002) based on publications between 1950 and 1990. According to the above, the objective was to propose new tracks classification of the coefficient of variation for the response variable for Brachiaria spp. Thus, CV% are tabulated-'s a response variable, experimental unit, and forage species in spreadsheets. We selected the most frequent variables in scientific articles indexed in Scielo, Directory Open Access Journals, Google Academics and Associação Brasileira de Zootecnia from 2000 to 2014. We tested the normality of the coefficients of variation with the Kolmogorov-Smirnov test modified by Lilliefors and Shapiro-Wilk, both the 5% significance level. We used the methods of obtaining Garcia classification tracks (1989) for CV% 's normal distribution, and the Costa et al. (2002) for all other with unknown distribution. For these CV's% sought to validate the tracks obtained by the method of Costa et al. (2002) using the chi-square adherence test corrected by Yates factor and when it was not used the Fisher's exact test, both at 5% significance. It is concluded that the coefficient of variation classification ranges for the variables crude protein, fiber, neutral detergent, acid detergent fiber, dry matter intake of pasture intake of total dry matter, leaf / culm, consumption of organic matter, dry matter intake of pasture intake of total dry matter, crude protein intake, fiber intake neutral detergent consumption of non-fiber carbohydrates, total digestible nutrients, grazing time, time rumination, and idle time obtained in this work can be recommended for the genus Brachiaria spp.
Descrição
Palavras-chave
Estatística , Gramínea , Planejamento experimental