Use of the Papadakis method to improve soybean trial quality

Abstract

The objective of this work was to verify the modifications in statistical analysis regarding the assumption fulfillment and the different precision indicators which can be attributed to the application of the Papadakis method, in the analysis of soybean grain yield experiments. The results of soybean yield of 226 trials of genotype competition, performed in a design of complete randomized blocks, were used. For each experiment, the assumptions were checked for: the variance analyses, hypothesis tests and other analysis to identify its precision. In order to verify the profit in precision, the same data were analyzed using the Papadakis method. The assumptions were adequate for both analysis types. The mean of Fasoulas differentiation index increased from 5 to 12.43, for homogeneous blocks, and to 13.85, for heterogeneous blocks. The determination coefficient, adequate to classify the precision of the experiments, was approximately 10% higher, when the Papadakis method was used, regardless the efficiency of block use. The efficiency of the Papadakis method does not depend on the efficiency of the blocks, and allows an improvement of the indicators of experimental precision.