Statistical Analysis, Use of Variety Trial Results

It is difficult if not impossible to attribute the performance of a hybrid/variety to data obtained in one season. So reported results are not absolute. Results should be judged as one of the tools to evaluate a hybrid.

It is good to also know that average yield is influenced by other variables that were hard to control.

Factors such as rainfall, temperatures, soil fertility, diseases, insects, and others were often assumed to have similar effects for each hybrid/variety, which may not be always accurate due to the interaction of factors.

Experimental design and statistical analysis are used to account for the effects of these factors assumed to affect each hybrid/cultivar similarly (random effect). To judge whether the observed difference between hybrids/cultivars is real or just a random effect, we use least significant difference (LSD).

LSD is used to judge the likelihood of whether the observed difference between hybrids/cultivars is real or not. LSD is a decision tool to evaluate the probability that the observed difference is due to chance or due to difference in hybrid/variety performance. We will like to stress to our readers that we assign (report) LSD only after we perform the analysis of variance and showed the null hypothesis (the hypothesis of no difference) is rejected. If the analysis of variance did not show significant at an alpha level of 0.05, we don't  run or report LSD value.

Probability levels range from 5% to 20% (expecting 80% to 95% of the time the observed difference is real) to declare significant difference. For on-farm trials like variety testing, 10 to 15% would be appropriate.

When two hybrids/varieties are compared and the difference between them is greater than the LSD, the entries are considered to be significantly different. Any difference between hybrids/varieties smaller than the LSD value will be considered non-significant.

Thus, the performance of all entries within the range of the LSD from the top performer is considered to be equal to the top performer. Under this condition the product expected market (economic) value attainable from the difference could be used to select among statistically non-significant hybrid/varieties. LSD value is a difference expected if all the plots are planted to the same hybrid/variety (i.e. the effect in performance one may achieve through crowding).

Publication of Results and Suggested Use

Results from variety testing are mainly reported through the variety testing Web page, the Fall Seed Guide and the Spring Seed Guide (corn and soybean) also known as Extension Circular 101 and 103 respectively.

Reported results should be considered as one of the tools to evaluate a hybrid/variety. Seasonal yields are influenced by variables that vary much across seasons. It is difficult if not impossible to evaluate the performance of a hybrid/variety from data obtained in one season or one location. Factors such as rainfall, temperatures, disease and insect pressure, and others often assumed to have similar effects for each hybrid/variety, may not always be correct due to the interaction of factors.