Findings, implications and lessons learned

In terms of data collection the case study has demonstrated

• how GenStat can be used to lay out the design of an experiment, randomise treatments to plots and prepare the recording sheet,
• how plot numbers were rearranged in alternate rows to allow the recorder to traverse the field experiment more easily.

From a statistical design and analysis point of view the case study has described the use and analysis of a factorial split-plot design and shown how different standard errors are calculated to compare means at different strata levels.

Split-plot designs have their place when it is difficult or impossible to apply all treatments at the same plot level. Thus, in this example, it would have been impracticable to plant landraces on different dates to individual plots within different rows.

Nevertheless, it is important to note that, with the larger whole plot size, whole-plot treatments will usually be estimated with a higher variance than sub-plot treatments. This needs to be taken into account in the experimental design.