Search

The statistical analysis of the results of studies on the effect of sowing time on the yield and quality of winter wheat grains, from the perspective of multivariate variance analysis and multiple comparison of averages, is carried out in the article. It was found that the application of the procedure of a multivariate generalized linear model allows comparing the variants of experiments for all indicators in the aggregate. An important advantage of multiple comparisons over paired Student comparisons is also the ability to isolate homogeneous subgroups of experiments. In this study, post hoc comparisons according to the Tukey criterion were applied, the results of which are presented in tables of homogeneous experiment subgroups and multiple comparisons of indicators’ differences. The distribution of experiment variants by zones of desirability of Harrington functions of yield and generalized indicator of grain quality, depending on shift of sowing time is also presented. It is justified that the maxima of yield and quality of winter wheat grains provide a sowing period of 10 days later than previously recommended for the region. The use of multivariate dispersion analysis and multiple comparison of average increases the reliability of statistical conclusions, allows you to distinguish homogeneous versions of experiments in agricultural production.