Minimum number of replications in maize hybrid competition experiments

Abstract

One of the problems faced when an experiment is planned is to determine the number of replications for each treatment. The number of replications is directly related to the results accuracy. The aim of this study was to evaluate the effect of the number of replications on maize hybrid competition experiments. The test was carried out in complete blocks with 12 replications. Grain yield was the variable analyzed. Based on those 12 replications, 4,083 new experiments were simulated. Each one of them was submitted to an analysis of variability and the average was compared by the use of the Tukey test. The proportion of experiments with significant differences and a minimum significant level under 5% increased directly with the enhancement in the number of replications, indicating at least 11 replications for this value not to be under 95% for grain yield. There was a convergence among the values of minimum meaningful difference (DMS) stabilizing around 10%. The stabilization of the DMS average for a larger number of replications indicates that it is not viable to get highly accurate tests considering only the increase in the number of replications.