Avaliação da técnica de Papadakis aplicada em métodos de estimação de tamanho de parcelas experimentais na cultura de soja

Abstract

The experimental plot size should be adequate to reduce the experimental error and increase the experimental precision, hence the need for more efficient experiments. And for the determination of experimental plot sizes many methods are found in the literature that present adequate results, but some methods have presented unsatisfactory results with excessively small or large sizes. To circumvent this problem, an alternative is the use of the Papadakis technique. Therefore, the objective of this work is to evaluate the use of the Papadakis method for size estimation of experimental plots with application in the soybean crop. This method consists of spatial adjustment using the moving averages of the calculated experimental errors among neighboring plots as covariates. For the plot size estimation, the methods of maximum modified curvature, method of the linear model with plateau and the method of maximum curvature of the coefficient of variation considering the original data and those adjusted by the Papadakis method were used. The Akaike information criterion, the adjusted coefficient of determination and the residual standard deviation were used to verify the efficiency of the methods. It was found that the use of the Papadakis method provides a good fit in the estimation of experimental plot sizes using estimates of the maximum modified curvature methods and the maximum coefficient of variation curvature method.