Statistical Methods And Validation Of Field Results
The manuscripts should report enough details of the experimental design so that the results can be judged for validity and the repeatability of the experiment is clearly possible i.e. the previous experiments may serve as a basis for the design of future experiments. The methods of statistical analysis must be adequately described, and references to statistical packages included in the text, and listed in the references section. Appropriate means separation procedure must be employed and explained. A statement of the number of replicates, their average value and some appropriate measure of variability is usually important. Information such as analysis of variance tables should only be given if they are relevant to the discussions. Comparison between means should be made appropriately, as failure of this may result in incorrect scientific conclusions. Common parameters used include the standard error of a difference between means (SED), or the standard errors of the means (SE or SEM) when these vary between means. A statement of significance level (P) attained in the differences should be included.
Field experiments that are sensitive to environmental interactions and in which the crop environment is not rigidly controlled or monitored, such as studies on crop yield and yield components, should be repeated over time or space, or both to demonstrate that similar results can or cannot be obtained in another environmental regime.
