Experimental precision related to plot sizes, numbers of treatments and replicates in forage turnip

Abstract

The objective of this work was to determine the experimental precision in the evaluation of forage turnip fresh matter weight, in scenarios formed by combinations of plot sizes and number of treatments and replicates. Data from fresh matter weight of 3,456 basic experimental units of 0.5×0.5 m (0.25 m²) were used. The scenarios (1,728) were planned and formed by the combinations of six plot sizes (0.25, 0.50, 1.00, 1.50, 2.00, and 2.25 m²), i treatments (i = 5, 6,..., 20), and r replicates (r = 3, 4,..., 20). For each scenario, 2,000 resamplings were carried out, with replacement. The average of these 2,000 estimates, in each scenario, was calculated for the least significant difference by Tukey’s test, the variation index, and the experimental coefficient of variation. These statistics, in that order, are adequate to evaluate the experimental precision. Plots with the size of six basic experimental units (1.50 m²) are sufficient to identify significant differences turnip fresh matter weight between treatments, less than or equal to 36.88% of the overall mean of the experiment, by Tukey’s test, at 5% probability, in experiments in a completely randomized design with 5 to 20 treatments and five replicates.