I led the UK Farm Scale Evaluations of GM crops, and I thought that was big. But this is REALLY big. The advantage of this study size size comes from the more local and detailed analyses, supporting the sustainable intensification of farming at the local scale without losing the national picture. I’d be surprised if it can be repeated anywhere else in the world, it relies on the availability of a lot of resources coupled with an integrated bureaucracy. Just conducting the experiment is a great achievement; getting the data in, properly validated and analysed is, to me, simply awesome.
To me, it also raises the question of the role of large scale experiments in agriculture. There is a perception that these experiments are not really needed, if we have intensive data recording over a large population of farms, coupled with some form of analytics to come up with some form of optimum management. I’m old fashioned, I tend to mistrust analytics. It’s very easy for algorithms to become misled by incomplete and biased data, it’s also very easy for them to fail to capture the key variables of value as farming and our environment changes. In other words, to me, even though the amount of data may be far larger than we have seen before, it is still subject to problems of bias and lack of completeness. But experiments were designed to address this problem; you decide what to measure and what to vary, and how to sample from a wider population without bias. This study supports my belief that experiments are still the mainstay of collecting evidence to support changes in farming. The challenge is to make sure the experiments are large enough, and conducted well enough, to be fit for purpose.