Christine Hackett

Modern cereal cultivars have usually been selected as high-yielding in official trials on soil whose layers have been deeply disturbed, known as inversion tillage. On farms, issues of cost and sustainability mean there is often less soil disturbance and other cultivars may be better adapted to yield highly under these conditions. In previous work on spring barley we identified some cultivars showing consistent differential adaptation to tillage. Here we analyse ten years of trials of winter barley under inversion and non-inversion tillage and show that the ranking of yields also differs with respect to tillage. This suggests there is scope to improve official trials and breeding selection trials by considering yields in non-inversion and lower-input conditions.

Christine's statistical explanation

The analysis of crop trials for significant differences in mean yield between cultivars is a classical area of statistics. The question here was whether the rankings of the yields differed significantly under two treatments, inversion and non-inversion tillage. A simulation approach was used, adding random variation to fitted values from a model of treatment + cultivar but no interaction. This gave the distribution of how much cultivar rankings changed by chance in data with no true differential effect of the treatments. The observed change in rankings of some cultivars in the experimental yields consistently exceeded that observed in the simulations, identifying cultivars with genuine rank changes.

Publications

Featured publications

All current publications