Crops Performance Test results available
By Todd D. Whitney
River Valley Extension Agent, Crops/Soils
Our new 2011 Kansas State Extension Fall Crops Performance Test results for corn, grain sorghum, and soybeans are now published and available through our K-State Agronomy Department at www.agronomy.ksu.edu/kscpt. These performance tests are useful to compare hybrids and as a possible to management tool as producers select hybrids for their next production year.
Statistics are used to interpret data, and certain terms are used in reference to the "statistical significance" of the yield test. For example, the CV percent or the co-efficient of variability can be used to estimate the degree of confidence that a producer might have that if he/she planted the same hybrids on their farm under similar growing conditions that the same hybrids would yield better than the other hybrids in the same test.
Generally, producers can be very confident in using the yield results when the CV (%) are less than 10.When the CV value is 10 to 15 percent, then the yield plot data is acceptable, but producers are encouraged to compare other yield locations as well. However, when plots have CV values above 15 percent, then Extension educators may question the confidence in the yield results being repeated on your farm. In some low yield cases, it might be the best information that a producer could use, but a researcher or educator may state that the results do not provide "statistical significance."
The coefficient of variation was created as a measure of population variation. Although many scientists may still use the CV to accept or reject the validity of trials, the CV is based on the assumption that the mean and error variance change together at a constant rate. So, most Extension specialists advise producers to rely on the LSD (0.05) values more than CV (%) for comparing hybrids.
The least significant difference LSD (0.05) provides a yield number for confidence comparisons. For example, if two hybrids have a LSD of 14. This means that unless two hybrids/varieties differ by more the 14 bushels per acre, there is little confidence that you can say that one hybrid will better than the other hybrid; even if the one hybrid had a higher yield in the study. Conversely, if a hybrid in the study yields more than 14 bushels more than other hybrids in the study, then you have confidence that (95 percent of the time) this hybrid will yield more than the other hybrids if the study is repeated (maybe on your farm).
More information with the crop performance results on the K-State Agronomy website at www.agronomy.ksu.edu/kscpt or www.ksre.ksu.edu or through the K-State Research and Extension - River Valley District offices (785) in Belleville (527-5084); Clay Center (632-5335); Concordia (243-8185) or Washington (325-2121).