“In January, I spent two days at the University of Nottingham re-learning statistics and field trial design. As a host of many field trials at Morley/TMAF I want to help make those experiments a success.
Statistics then and now
Years ago, when I studied agriculture at Harper Adams, one of the options my friend and I chose was field experimentation. I remember it being maths-heavy as we had to work out the all-important coefficient of variance by hand! I hasten to add that although it was not in the days before computers, they were only used on special occasions.
Twenty-five years later, here I am managing a farm hosting field experiments. Whilst I don’t do the trials myself, I do have a large say in where and how they are located on the farm. As host, I want to help make the experiments as good as they can be. This is why I jumped at the chance for two days learning about statistics and experimental design in fields.
On course for better trial work
The course was led by Professor Debbie Sparkes and part-funded by TMAF for thirteen agri-professionals from AHDB, NIAB, BBRO, LEAF, G’s and Agrii, as well as TMAF. We all work with trials but not always doing statistical analysis. Professor Sparkes explained the importance of statistical analysis. It shows the likelihood of what is being tested being a true difference rather than just chance. This sounds straight forward. But how do you choose the right number of replicates, plot sizes, field orientation? What if you are testing multiple treatments for example, 3 varieties with 3 different cultivation methods and with or without irrigation?
Many field trials with varieties, fungicides etc. are 2m x 9m plots. The methodology is well-proven, the analysis of data is a routine operation. Much of the applied research of late is to do with soil characteristics, use of cover crops, cultivation techniques, rotations etc. This type of work is very difficult to do on small plots. This presents problems with the practicalities. How can it be done with larger scale equipment? How much space need it take up in the field? These are the discussions and decisions that have to happen.
Team work to make field experiments a success
It seems to me that there can be a lack of communication between the people who design experiments and the people fitting those experiments into the fields. My strengths are understanding crop growth and practical management we do, such as cultivations.
By attending this training, I now know more about statistical analysis and also have an appreciation of how to test treatments scientifically in a field. The course built my confidence to challenge researchers on their trial design and statistical analysis. It also highlighted how to present data to give the audience confidence in the experimental results.
All of that will improve the chances of the TMAF-hosted field trails yielding results useful to farmers.”
Stephen Aldis, BBRO mechanisation researcher recounts what he gained from the course.
“Having been with the BBRO just over a year, and with our field experiments being quite different from mechanical testing I have done previously, this course was a great opportunity to get a greater understanding of how to approach field trials with statistics in mind.
Any trial undertaken obviously sets out to achieve a scientifically significant result, whether it be positive or negative. Professor Sparkes talked us through the basics of the calculations used. This clearly highlighted how trial design can affect the chance of achieving the desired significance result.
The course provided a fantastic introduction for me to have greater input and understanding of the trial work I am involved with and the results. It has also equipped me to interpret and question data generated & presented by others.
It was fantastic to participate with such a diverse group of people working in and around agricultural research. Understanding the issues they face, and how they overcome them, was definitely an added bonus.”