The magnitute of the experimental error has a direct effect on the success of plant breeding experiments. This study aimed at evaluating the interference of stand correction methods on the yield of 25 grain sorghum hybrids, in seven environments. The methods tested were: without correction, rule of three, covariance of medium and ideal stand, Zuber, Cruz, Vencovsky & Cruz and stratified correction, based on genotype groups. The Pearson correlation coefficient was used to assess the magnitude of the influence exerted by each method, concerning the data without correction. The stand correction methods based on covariance and Vencovsky & Cruz reduced the coefficient of variation and showed a high correlation with the adaptability and stability parameters of hybrids, however, these differences were not so significant to justify such adjustment.
Sorghum bicolor (L.) Moench; statistic analysis; experimental error
