Statistical Analyses and Interpretation

Data collected on all traits measured during the 2006-2007 crop season are presented in Tables 5 through 14. Data presented for individual locations were analyzed using a lattice design. Regional, state-wide, and multi-year data result from analyses based on a randomized complete block design. If an observation was missing in one replication an adjusted average of the remaining observations (least squares mean) was used to approximate the missing observation. Mean comparisons were made using Fisher’s protected least significant difference (LSD) at P=0.05. Coefficients of variation (CV %) were calculated from the analyses of variance of each location and across all locations.

The LSD is used to compare the performance of two specific varieties. If the mean of a variety exceeds that of another variety by more than the LSD, then the difference observed will be a true difference in 19 out of 20 instances under conditions similar to those of the test.

Table 14 ranks soft red winter wheats according to their state-wide average. Overall rank can be very misleading. Growers should be careful to make pair-wise comparisons of results from both the appropriate location or locations and the state-wide averages before selecting one wheat variety over another for production in Missouri. Variety selection should be based on yield stability in a production environment over years and locations. Where a variety has been in the test for two or three years, combined analyses of the yield data over years are presented. In choosing a variety, other characteristics such as test weight, heading date, and disease resistance should also be taken into consideration. Where disease data were not reported in a particular production environment, they can be evaluated from locations in which they were rated.