Meta-analysis is a powerful tool to summarize variety mixture effects - exemplified by grain yield and weed suppression of spring barley

Summary

In a new project, we aim to increase the general understanding of the power of meta-analysis to combine existing experimental results on variety mixtures. In this way, explanatory power can be increased compared to separate analyses and overall measures and relationships may be revealed. We will thus pursue to uncover a number of critical issues, including the relative importance of various traits and trait combinations on mixing success. To demonstrate the methods of meta-analysis applicable for variety mixture data, we considered a data set consisting of grain yield and weed ground cover assessments in 16 field trials of six 3-component variety mixtures and their components (part of the Danish BAR-OF field trials). The effects of mixing were analysed separately for each field trial and the results used in a meta-analysis in combination with their standard errors. We also analysed the mixing effects of each mixture by fitting a linear model to the entire data set. Both methods showed an overall positive mixing effect on grain yield and a trend for less weed to be found in variety mixtures. Finally, strengths and shortcomings of the methods are highlighted.