174 – Twenty Years of RMS for CRFS: Introducing Roger

Lucie: [00:00:00] Hi and welcome to the IDEMS podcast. My name is Lucie Hazelgrove Planel, I’m a Social Impact Scientist and anthropologist, and I’m very pleased to be here today with Roger Stern, with whom we’re going to record several episodes and all with a focus on research methods for agriculture.

Hi Roger, and welcome to the podcast.

Roger: Hello, Lucie. It’s nice to be here.

Lucie: I think it’s your first recording with the IDEMS podcast, isn’t it?

Roger: It is my first IDEMS podcast, so I’m a beginner.

Lucie: Very exciting.

Now you were part of the Research Method Support team for the first 10 years or so, I think, of what was at the time the Collaborative Crops Research Program, which is now the Global Collaboration for Resilient Food Systems. I’m very curious then to know how that happened, basically. How did you get to being part of the Research Method Support team? [00:01:00] What’s your background?

Roger: My background is what you might term a research method supporter. Before CCRP, I knew Niger quite well because I worked there for one of the international centers called ICRISAT for seven years in the 1990s. And… 

Lucie: Sorry, Roger, what were you doing? What were you doing at ICRISAT?

Roger: I was research support. I was the statistician responsible for the different ICRISAT centers throughout Africa, there were six of them. And I was helping people with their analyses and also helping before the season with discussions on the design of the studies for the forthcoming season. So I was providing what I would call statistical support.

And one reason I was keen to be based in Niger was that Niamey is a curious [00:02:00] center and one thing it combines is ICRISAT being there with also two regional climate centers. And I have for a long time been very interested in the role of climatic data in agriculture and more generally the analysis of climatic data. So that was a good reason for being based in Niamey.

Lucie: When was that? You said it was in the nineties.

Roger: This was from 1990 to 97, many years ago. And, just to position it, when I arrived agricultural research, particularly crop breeding, was very much an on station exercise. And by the time I left the move had started, and started rather poorly, that much more should be done on farm. And that has moved on and not looked back.

Lucie: Yeah.

Roger: That’s been quite an interesting transition since I left [00:03:00] but not since I was involved in the agricultural research support. Coming back to CCRP, in West Africa the first meeting was 2006.

Lucie: Yeah, exactly. So what you doing between…? 

Roger: 97 and 2006. I was at the University of Reading in the UK, but I was in a self financing center basically working as a statistical consultant on jobs doing roughly the sort of support I was doing at ICRISAT, but much broader, not just for agriculture, where we got consultancy work to help people on the design of their studies, be they medical or industrial or agricultural, and on the analysis of data generally.

I’ve tended now recently to call myself a data scientist because statisticians are rather unpopular species and people say they don’t like statistics. They [00:04:00] don’t seem to be so phased by data science and the overlap is sufficiently great. I like to think of myself as a data scientist now. 

Lucie: Well, it’s interesting, you mentioned that when you were at ICRISAT even you were supporting people design their trials. And to me, I don’t always see that as being something that a statistician would do. It feels like a data scientist is more general and so they would be able to support in these other things. 

Roger: That’s interesting you say that. I see one of the distinctions between a statistician and a data scientist is that a data scientist often starts with the data, whereas the statistician, if they’re a good statistician, starts with how to collect the data and should be advising on that.

In fact, one of the things that many consultees don’t seem to realise is that they’ve often been taught statistics rather badly. So they’ve only been taught bits of analysis and the things they really dislike. And [00:05:00] when they show their data to a statistician, statisticians traditionally ring their hands and say, you should have come to me earlier and I could have helped you collect the data better.

But the secret of many statisticians is that because nobody comes early, they don’t actually know what to do if you did come early. I would claim that after my time at ICRISAT, I did know what to do. I was able to help people design a survey, design an experiment or field trial, and it’s a very important part of statistics. Whether the design stage is recognised to be important for data scientists, I’m not clear.

Lucie: Interesting. Great.

So you were at Reading University, you were providing lots of support to different people and then the CCRP came up, or what happened? 

Roger: The first meeting of CCRP was in 2006 and it started with a community of practice [00:06:00] meeting, and I think they continue, a one week meeting, and I was invited to the first one.

Lucie: And this in West Africa, I think.

Roger: This was in Niamey and I was invited because they wanted to understand right from the start the role that climatic data could play in a research program. So I was invited to give a day on climatic data in agricultural research, if you like. Which I did. And I have to say it’s now 20 years since CCRP began, and nothing much has changed on climate.

We had another interesting initiative six years later when I was fully involved in the research support and the farmers organisations said not knowing about climate variability was one of their failings, and we then try to do something else. But really it’s been kept apart from the research as it is for a lot [00:07:00] of agricultural research. 

Lucie: I was going to ask, is that a general agricultural research problem? 

Roger: I think so. And many years earlier, I spent a year in Nigeria where I was asked to learn about climatic data because the vagaries of rainfall are one of the main reasons for variability of yields in agricultural field trials. And so you would think that something which causes a lot of variability would be studied quite a lot, and it isn’t. But nevermind. People are much more interested in studying the things they can control than the things they have to put up with.

Lucie: Yes. 

Roger: Continuing my story, I gave my one day, which has had very little influence.

Lucie: Just one day! 

Roger: It was one day within the week. So it was a very nice opportunity right from the start. And it didn’t work out very much.

Lucie: When you say one day within the week, in the community of practice in West Africa we have [00:08:00] a big community of practice meeting every year which involves many different activities. And it sounds like that’s the week you’re talking about then.

Roger: That is exactly the week, and often I think you have themes in particular community of practice. So mine was a theme in the first one.

Lucie: Exactly, okay. 

Roger: And in 2008, a researcher in Niamey, who I think was now a grantee from CCRP, called Bettina Haussmann, who continues to be very well known, to, the program in West Africa.

Lucie: Well, she’s currently the Scientific Lead, and has been I think for quite a while of the program, the community of practice in West Africa.

Roger: That’s right. She remained as a millet breeder in West Africa till about 2000, I dunno, about 2011. And then she went back to Hohenheim and very quickly took over [00:09:00] her role in what was then CCRP. And I think it suited her very well, and she was brilliant at the role, so I’m not surprised that she’s continued. 

Lucie: Yeah.

Roger: So she asked whether we could give the researchers in the young CCRP some training in statistics because that was demanded.

Lucie: After this workshop. 

Roger: Yes, this was, I think in 2008, and I gave a one week training course in Burkina Faso to various invitees from the three countries, Mali, Burkina, and Niger. And I have to say at the same time, in our center from Reading, Carlos Barahona, was asked to give a similar workshop to CCRP in the Andes.

Lucie: Ah, okay. Interesting.

Roger: And so we gave those two workshops at the same sort of [00:10:00] time. And this caused a surprise to the powers at be, the management of CCRP. And the surprise was that the participants had quite liked receiving their statistics course. I think the management of CCRP was amazed that anybody would like statistics. And the grantees had certainly not liked much statistics before. And I would claim, I think that this was the beginning of discussions that the Statistical Services Centre at Reading University and the CCRP management had which led to the role that now continues with the organisations that have grown from Reading providing research methods support.

Lucie: Exactly. So our current partners, as IDEMS, our current partners in providing research methods support are Stats4SD, Statistics for [00:11:00] Sustainable Development, who Carlos Barahona is the current director I think of.

Roger: That’s correct. And and Stats4SD, like IDEMS, spun out of SSC, which was totally within the University of Reading. And so these were two spin outs, which continue. And between them, they handle this research method support for, you must tell me what CCRP has become because it was CCRP when I left it.

Lucie: The Global Collaboration for Resilient Food Systems or CRFS for short.

Roger: So, my involvement in the early years was very much on the mixture of attending the inception meetings, which was then what happened after a grant was given to try and conceptualise the research. 

Lucie: Yeah, and these meetings are really important even now, it [00:12:00] enables partners to come together to plan what they’re going to do, to make sure that they’re all in agreement.

Roger: Yes. And I think from the beginning, and I think it’s continuing, I found that at those meetings the grantees were sometimes feeling slightly uncomfortable that the grant was taking them outside their comfort zone, particularly in the collaborative work they were being asked to do. But also in some of the activities that they designed. And I think that that was always an exciting part of the work.

So we then were involved very much in the analysis of the data that came. I have to say that I understand now things have moved quite a lot. There were not so many students in my day who were central to the research. And so we are often working with the [00:13:00] researchers from INERA in Burkina from IER in Mali and from INRAN in Niger, as well as from the university. So we were often working with the staff, and there would be some work for a PhD student, but I understand it’s very much more a student activity in the more recent grants, or they’re more central maybe. 

Lucie: It depends on the projects. Some of the really older projects who started back when CCRP was created, and there’s even been this other podcast episodes about Professor Baoua, for example, telling how he’d developed this whole huge team across the three countries, Mali, Burkina Faso, and Niger using this human wealth of students and of PhD students supporting masters or undergraduates.

So there’s a big wealth of students definitely in the CRFS as it is now, and they are very much encouraged too, within the community, as they are the researchers of tomorrow.

Roger: That’s right. And people like [00:14:00] Baoua were in my day, when I was looking after it, Baoua and also on Food Systems, Moustapha in Niger as well. Baoua is FUMA Gaskiya, is that correct?

Lucie: No, he is the lead for Sahel IPM, the Integrated Pest Management project, which used to be called GIMEM I think, because it wasn’t looking at all pests. It was specific for one. 

Roger: It’s still what I was involved in. And I think that’s one of the features of the research program that’s been very welcome, that often the donors come in for three years and then they leave. The fact that when important researchers are doing key work, this donor, McKnight, has continued with them to foster this research for quite a long period. It is very welcome.

Lucie: And the results are incredible that have come out, we’ve just mentioned Professor Baoua and Moustapha Moussa. The results of their projects are incredible.

Roger: Yes.

Lucie: It speaks for itself, really.

Roger: Yes, that’s right.

[00:15:00] I was going to mention on the research support, I found compared to many other agricultural projects this research support component was a feature of the McKnight support, which was not done in many other projects that were often on a much bigger scale.

So I found that many of the grantees were learning quite a lot of skills both from the capacity building that was part of this as well as support for the analysis of their particular data set that they then were able to apply to their other research. So the grants were relatively small, but seemed to be very much appreciated by many of the grantees.

Lucie: This is always what is so interesting to me that all researchers need research methods support really, and in your own work at University of Reading it also demonstrates that. And so it’s very strange that there’s so few, so few [00:16:00] opportunities for support available.

And as you’ve mentioned, the fact that these people are using what they’re learning, the skills they’re learning, the ways of thinking that they’re learning in their other work outside of the McKnight Foundation’s programme. It’s so clear. Again, the benefits are so clear of providing a support role for researchers, whether it’s a research project in fact, or any other development type project.

Roger: I should mention that with students, I have a principle, however, that it is the student that is going to get the degree. Therefore, the student must be involved in all stages of the work. The research support does not mean we analyse data for the student. We analyse data with the student so the student can write up the analyses and they may not have been able to conceptualise it all at the beginning, but they can understand it all and therefore understand the results at the end. So the degree they get is genuinely theirs.

There are some instances where I think a [00:17:00] student would like us to take over and then I would like that degree. I don’t see why they should get the degree if we were doing all the work. So we helped them and we guided them. But particularly when students are getting a higher degree, the higher degree reflects their own work, and therefore they did understand the analysis. And I think this helped them to understand their trials and their surveys much better.

Lucie: Exactly. And that capacity building aspect is really central to what we do now in terms of our research method support. We are always trying to think about how, especially for those who are still in their learning journey, like very clearly, obtaining degrees, how we can support them in having the skills to when they become lecturers and researchers, fully blown lecturers, what skills they need in that.

And I think that’s a really interesting point you’ve made there about what it means to give research methods support. Are we just doing the work? How does that balance of activity come in? I [00:18:00] agree that it is a bit different when it’s a project lead, for example, who has set up a huge project and don’t necessarily have the time to do all of the analysis themselves.

Roger: That’s correct. Yes. You were also saying you wanted to know how things have changed. 

Lucie: Have things changed? Well, no, sorry, I know things have changed. Did things change in the sort of 10 years or so when you were part of the CRFS?

Roger: I think the big change was the extent to which the farmers organisations were involved.

Lucie: You witnessed that, you were present at that, okay.

Roger: That started, I think Bettina was very instrumental in that change, and it was during my time, and towards the end of my time, was I think the first time that the farmers organisations, they were involved. The question is as a farmer’s organisation, should they do research?

And I think there’s often confusion between development and research. And if development involves [00:19:00] innovation, which is very much what farmers organisations are trying to do, then that is research. So I didn’t see this big distinction, and I think the big change was the extent to which they were involved as full partners, rather than being, if you’re like, subcontracted always by the national research program as facilitating the work but not involved in the research components. And that big change was happening even in the early years of CCRP.

What didn’t happen until after I left was the scale of some of the on-farm trials which I think is facilitated by the farmers organisations being senior partners. When I was involved in on-farm experiments, they were less satisfactory because they only involved a [00:20:00] few farmers, maybe 30. But when you think of the variability in farming practices from farmer to farmer, 30 is, if each farmer has one replicate of some treatments, 30 sounds quite a lot. But in terms of the on-farm experiments, the farmers’ practices were sufficiently different that 30 was rather small.

And this fitted in with the new ideas of Options by Context where the context for different farmers was sufficiently different that one actually didn’t want to learn about farmers as a whole. Your interest was not what was the best treatment, but what were effective options for different farmers or for different types of farming.

Lucie: Exactly, and the environment they’re in, the sort of soil types, perhaps the household size, it’s the household’s interest that they want to [00:21:00] maximize one thing over another, all of those sorts of contexts.

Roger: Yes, that’s right. Could they get a loan? Could they afford fertiliser? Some farmers actually worked as labourers at the beginning of the season. Therefore, they couldn’t do early planting on their own fields. So there were all sorts of different pressures on the different types of farmer, and one wanted to have solutions that suited as many as possible.

Lucie: And we’ll no doubt talk more about Options by Context at another moment.

I did want, I was just thinking though, that at the moment we are currently, not only in terms of research method support, but also the wider world, we are all wondering about this recent technological change of what to do with artificial intelligence and how to make the most of it in a responsible and intelligent way.

But I think when you were, back in 2006, were people using computers to do their analyses as much? I’m [00:22:00] wondering if there was technological changes even when you were providing research methods support for the CCRP.

Roger: Well, I think the big change early on was that in my early years people would say, oh, you come from England, you have lots of data there. We don’t have any data from where we are. So they used to think of their trial or their survey as being the first attempt. By the time CCRP, this was in the nineties and going on a bit longer, the rise of data science as a discipline, tied in with the realisation by many developing countries that the world was awash with data.

But I think many people in the early years of CCRP had not recognised that sufficiently, that they spent enough time on the work that related to their data that provided a good background. And now I think, if you like, [00:23:00] the artificial intelligence community now is putting all that work together.

And so if I see a role for it, it’s to provide background information to be able to summarise what is known in a highly efficient way, and therefore to see what is not known and on which you might do a survey or an experiment to compliment.

Lucie: Oh, it’s not very good at that because of data issues, a lot of the AI algorithms don’t have access to that sort of data.

Roger: That’s correct. There’s still a lot of scope, for example, for national surveys, which are often sparing in the way they share their data to become important as background information within a country because the researchers can get hold of that. So that’s correct.

Lucie: But most, you’re talking about surveys there, most surveys when you were part of the CCRP were no doubt on paper.

Roger: That’s correct. The big change, which really happened after I finished, [00:24:00] even when I was doing that, it was quite difficult collecting data electronically for a survey. And the tremendous advantages of ODK, of being able to collect the data electronically, therefore, to be able to check your results, your data instantly, and then to move quickly to analyses are just wonderful.

You still have to be very careful, the data you collect electronically is carefully collected and good quality, but the tremendous benefits of not having to computerise the data is just wonderful.

Lucie: There are some problems with it too, in the sense that when people collect, data on a computer or on a phone, they tend to give less information. For example, I’m an anthropologist, so I’m more interested in the longer answers where people elaborate a bit more, and using a tool like ODK or another sort of online or not necessarily [00:25:00] connected to the internet, but another digital survey type tool, it often doesn’t get that sort of depth or doesn’t enable people to share as much.

But I think there are ways of using the tools, even just by using the recorder, the voice recorder aspect of the tools to get that same information, which is interesting.

Roger: I would argue that I agree with that and I’ve seen many recent surveys where some people have not given an answer or given a one word answer, while others have given a paragraph. I think I would claim those examples, the failings are much more with the team collecting the data and the way they ask the questions. It’s very easy to blame the data collection tool.

Lucie: Absolutely.

Roger: Having stories that are paragraphs is very important and compliments the summaries, and they should be asked in slightly different ways.

Lucie: Well, we’re in, in agreement [00:26:00] there.

Roger, this has been fascinating. We will continue our conversations in future episodes. It’s been a pleasure. I have many more questions for you and I really look forward to hearing more of your stories and more of the changes, more of the depth of what you were working on.

Roger: I look forward to them too. Thank you for now.

Lucie: Thank you.