Blog

Moving Science in Practice – What the Statistician Learned About Agroecology

The area of agroecology is a key focus of the work we do at Stats4SD, particularly through our ongoing projects with the Collaborative Crop Research Programme (CCRP). However, as someone with a background in statistics, focusing more on medicine and economics, this was something I had learned about entirely ‘on-the-job’ rather than through any formal introductions.

As part of a new partnership we have established with the Agroecology Learning Centre at the University of Vermont, I was invited to be a student on a week-long course to learn more, and dig beyond my very surface level understanding of the issue. The course involved spending some time learning about the background and history of agroecology, followed by a week moving around various farms in Vermont, to see examples of agroecology in practice. This really helped to start piecing together and expanding on the various isolated fragments that I had previously encountered.

So, here are my top three take away-tips about what agroecology is, and the role of research and statistics within the context of agroecology:

1. Nobody agrees what agroecology is (and that's fine)

The shorthand I tend to go with is that agroecology is ‘agriculture with the intention of maximising the benefit to the surrounding environment and ecosystem’. By comparison, the ‘traditional’ agriculture model is generally thinking instead about maximising the crop yield. But unlike other alternatives to traditional agriculture, agroecology takes in a pretty broad definition of environment. This includes not just ecological factors, through encouraging biodiversity and improving soil health, but also social factors, through ownership, labour conditions, supply chains, and nutrition.

Many organisations have set about creating “principles” (CIDSE) or “elements” (FAO) – to help define and guide agroecology. Although how these are put into practice will be interpreted differently depending on specific needs.

Each of the farms we visited in Vermont took a completely different approach to applying agroecology. The Cat farm was primarily run as a farmer training school, looking to inform and inspire a new generation of farmers to follow more sustainable and ecological practices. The Bread and Butter farm embraced small scale experimentation, constantly looking for and testing out the impact of applying new practices. The Digger’s Mirth farm were extremely worker focused, trying to balance a collective organisation of labour with a productive and ecologically sound farm. And the Farm Between were most focused on ways to steward the land and restore biodiversity and environmental health.

 2. Contextualising these differences is key to success

For successful agroecology, it is not just a case of changing from one set of recommendations to another, but instead working out exactly what needs to be done differently and how to best meet the target goals. Specific local niches and requirements will result in different ways of working out how the farm production will be set up to fit the different personal priorities of the farmer.

For traditional large-scale agriculture, the end goal of production is the same across different regions – increased yields. But, when considering agroecology, there is no singular goal. The elements of the farming system should not be dictated down through research. However, through research done well, recommendations and knowledge about what has worked in similar contexts is vital to be able to start making informed decisions about exactly what and how things should be done differently.

3. Research and analysis should be matched appropriately to the system

With the added complexity of the systems being created, and of the range of desired outcomes, comes an increased need for tailored research designs and analysis. Research helps us to understand whether changes made to the farm are working as expected; how much they are working and allow investigation into how they might be expected to work in other contexts.

In traditional agricultural research, new innovations are closely analysed in replicated conditions, with randomisation and controls. But replication, randomisation and control all - to a certain degree - are antithetical to agroecology. No two farms are true replicates; agroecology interventions should be purposively chosen to meet the targeted goals, and placing controls into research provides an environment which is not a true reflection of the real world. This is not to say that these are ideas that should be discarded completely, nonetheless the research itself needs to complement the complexity of the system - rather than ignore or restrict it. We talk a lot in Research Method Support (RMS) about investigating differing “Options by Context” (OxC), and this idea is key in being able to derive meaningful insights from research of complex systems.

Spending time in Vermont at this workshop provided a completely different insight into agroecology from that which I had been exposed to before predominantly working with CCRP projects in East and Southern Africa. Although the farms, and resources available, are almost incomparable between these two contexts, there were principles and ethos of more sustainable and equitable farming practices in common between the two. Examining agroecology within this different context has given me a much better understanding of just how broad this area can be, and how many different elements can be adapted into an agroecological framework.

John from The Farm Between in Vermont, who sees his role in agriculture as a ‘steward of the land’

Dairy farming is the #1 source of agricultural income in Vermont. Stoney Pond farm did not consider themselves to be an agroecological farm, but one which incorporated some agroecological principles into their work.

Me, with some of the other course participants seriously contemplating the question behind us at the Farm Between.

Author: Sam Dumble

Sam has been part of SSD since its inception in 2016. Given the importance and frequent misinterpretation of statistics, he is keen to ensure that results are presented in a clear and understandable manner, particularly through the use of graphics and maps. He does most of his analysis using the software packages R and QGIS. He is also an experienced user of various other packages such as SAS, SPSS, Minitab, Genstat, Stata and CSPro.