201 – Reflections on the World Statistics Congress

[00:00:06] David: Hi, and welcome to the IDEMS podcast. I’m David Stern, a founding director of IDEMS, and it’s my pleasure to be here today with James Kaleli Musyoka, an Impact Activation Postdoctoral Fellow. And we’ve just come back from the World Statistics Congress.

So today I’m hoping to dig in and discuss what that meant to you. It’s not your first.

[00:00:30] James: My second.

[00:00:32] David: Yeah, the first one was back in something like 2017.

[00:00:36] James: Yeah. In Morocco.

[00:00:37] David: Yes, that’s right. That was a big deal, of course, because this was one of the rare times that it’s been on the African continent, so we made a big effort to get people there.

[00:00:48] James: Yeah. We were a big team there.

[00:00:50] David: Yeah. And that’s where you sort of launched the African Data Initiative.

[00:00:54] James: Yes. I think that’s the other reason why it was special to us. We launched the African Data Initiative and yeah, we had a lot of presentations on that and R-Instat.

[00:01:06] David: Yeah. And here, eight years later, you found yourself in the Netherlands at The Hague. Why is this an important event for you?

[00:01:17] James: Interesting question. So I think, to me the most important thing from this event is, what the event is about is it’s bringing together statisticians, or people working in the, you know, the whole statistics community together. So it’s an opportunity to share what we are doing. And also to listen to others, you know, get to know what they’re also doing.

And I think we also get to interact and build collaborations where we see potential people that we could collaborate with. I think that’s the biggest opportunity that I’ve been, you know, been able to seize from this meeting.

[00:01:56] David: And for you in particular , this sort of goes alongside ICOTS, which is the International Conference on Teaching Statistics.

[00:02:03] James: Yeah.

[00:02:05] David: Which is also part of the same community. You were at a round table related to IASE and then you became a vice president of IASE.

[00:02:16] James: Yes. That was, I think, just before COVID until 2023. And ISI, you know, comprises these other associations like IASE, which I think that is like the main organisation that I’ve been part of.

[00:02:31] David: The International Association of Statistical Education.

[00:02:34] James: Yes.

[00:02:35] David: And both of us have served on that as vice presidents, we’ve both been involved, actually for the same amount of time, because we attended the first, our first ICOTS together, back in 2010, was it?

[00:02:51] James: Yeah, that was my first, I think, time interacting, being part of this community, actually attending an international conference.

[00:02:59] David: Unless I’m mistaken, it was your first time to Europe, a lot of firsts in that trip.

[00:03:05] James: Yeah, exactly.

[00:03:06] David: I suppose to keep going on, why is statistics in a community of statisticians still important in the era of big data and data science? Should this, you know, I don’t believe it should, but why is a community of statisticians when everybody’s talking about AI and data science, …

[00:03:29] James: Machine learning.

[00:03:30] David: Machine learning. Why a community of statisticians? Why the World Statistics Congress?

[00:03:37] James: That’s a very interesting question, again. That reminds me of the keynote speaker during the 2017 World Statistics Congress, where I think it was Hadley Wickham, I think he was being asked, you know, is data science, just statistics and all that?

I think statistics is still an important subject, I think there has been some modernisations of methods. We see data science, machine learning, and actually I think in the Congress I saw a lot of sessions on data science and machine learning. So I think statistics is still very important in the world that we live in. And there are a lot of relationship between statistics and of course data science and machine learning and all that.

There used to be a big gap between the two, but now I think statistics is also undergoing modernisation and I think a lot of statisticians are now sort of data scientists in a way. So I’m struggling to see whether, you know, there’s a difference between the two.

[00:04:36] David: Well, let me come on this because you brought up Hadley Wickham’s talk and he had this wonderful expression, is data science part of statistics? And his answer was, it could have been, but that ship has sailed. And that was back in 2017 and it sailed even further now. So these are two genuinely different tracks and the importance of statistics is still there.

[00:04:59] James: It is still there.

[00:05:01] David: But broadly, the way I like to distinguish them is, statistics was born in data scarcity and data science was born out of data abundance.

[00:05:13] James: Big data.

[00:05:14] David: You know, when you have so much data that your traditional statistics methods are not answering the questions that you care about.

And we still need both. We still need to gain knowledge in data scarcity, and there are rigorous ways to do that, which use the foundations of statistics, of design, and the methods and statistics to make the most out of a small amount of data.

[00:05:38] James: But the reason I was saying that I think that, it’s interesting you say that the two fields are still different sort of, but don’t we live in a world of, you know, full of data? We are in the data revolution age, and I think that has somehow influenced how the statistics subject is now being taught.

We had several talks, which, you know, were touching on this, that now we are bringing in data into the teaching of statistics, real world data. And so I was thinking that we are revolutionising statistics education because we are living in the age of big data, and the two getting closer.

[00:06:21] David: Well, this is the beautiful quote from Hadley Wickham. They could have been the same, except statistics defined itself as coming out of data scarcity. And therefore, it allowed data science to get away and now be defined as coming out of data abundance. In the world, and within this data revolution, we need both.

And yes, they should be coming together, and yes they should be taught more alongside one another to understand what one brings, what the other brings, how within data scarcity, you can remove biases in certain way and how big data can solve problems that you couldn’t solve before with small amounts of data, but where you need to be careful because of the biases which you cannot control.

So there’s definite differences which exist, whether you have a mindset of data scarcity or a mindset of data abundance. And it’s not that they are fundamentally different, they could have been under a single umbrella. And arguably they should still come back together under a single umbrella.

When I was young, I always thought statistics was the science of data, which does sound awfully like data science. But in some sense, now in our communities, in our society, there is value to having a community which is still coming out and leading out from the knowledge that working in data scarcity brings and approaching big data problems and integrating in with data science.

But the World Statistics Congress, to me, represents a mature perspective on data in ways that I feel data science, and, you know, you see this with recent advances in generative AI and other uses of big data, they’re still in their infancy, they’re not mature yet.

So actually it’s not that statistics and data science shouldn’t be coming together more as you are advocating, and I agree, it’s that they are at different levels of maturity. This is like an elder who doesn’t have the same energy as a youth having to work together.

[00:08:49] James: Statistics is the elder here.

[00:08:51] David: Statistics is the elder, data science is the youth that brings energy and enthusiasm, but maybe not quite the same sense of responsibility and carefulness and care and attention, not the same level of experience. And that image has been used so often.

But it’s often been used presenting statistics in a negative light, and I don’t think that’s the case. There’s so much, and I found this actually in this event, there’s so much value that comes from recognising, maybe slowing down a little bit, recognising we do need to be careful, we know about some of the challenges that come when you are handling data. If you are rushing with data, you can be misled.

[00:09:42] James: Exactly, yeah. There were a few sessions in the congress relating to data science, machine learning, and also the education side of those sectors.

[00:09:50] David: And I think we should do another session with Lily on our own sessions that we had there and what they were trying to communicate. Because I think there was some really nice points that came out and they were very well received, actually. And a lot of this comes out of the work you started on the African Data Initiative, and so it is all broadly within that remit.

But to keep a focus on the value of ISI and you know, the World Statistics Congress, the International Association for Statistical Education, these communities have experience and knowledge. What I would argue they’ve failed to do, which data science has done amazingly, is captured the public’s imagination to really get people interested in data.

And I would argue that the challenge that’s facing these communities and, I suppose, other communities in statistics and data, and you are a member of the Finnish statistical society, I’m a member of the Royal Statistical Society. These communities centered on statistics, I believe can be bringing their expertise, their experiences to bear in the gold rush that is data science.

[00:11:23] James: Yeah, just thinking about IASE, the one for statistics education. I think it’s, in my view, I think it’s more concerned or involved in the education of statistics, which now we are seeing several initiatives of people getting, bringing in data into their teaching of statistics.

So I think their focus has been mainly on, you know, the pedagogy, how we teach statistics basically, to improve the teaching of statistics. And I thought, for that organisation, I thought that they’re sticking to their mandates, really? Do we expect them to like…?

[00:11:57] David: It’s interesting. I would argue that they aren’t, they are expanding their mandate to include the education of data science at all levels. I’ve known many people within IASE who are involved in data science education and are bridging the gap between statistics and data science, and trying to see, well, what is it that we need to keep from the statistical training? What is it we need to improve?

We know very well there’s not enough work on cleaning data. This is something which we’ve argued and put in in a lot of our work, and others in the community are similar. Actually changing the nature of statistical education so that statistical education and data science education run alongside one another.

There’s some very nice examples from within the community where they have successfully attracted joint programs, which have statisticians and data scientists working together as part of a postgraduate program, which actually means you get really well trained both statisticians and data scientists because they have an eye on both worlds.

Some wonderful work happening in Australia and New Zealand on this. They really are, in my mind, at the forefront of what’s possible and what’s going on if you’re thoughtful about it. And just as a sort of small plug, next year, ICOTS is in Australia. So if you want to find out about this, it’s the perfect opportunity.

[00:13:29] James: Yeah.

[00:13:31] David: It’s the conference where I got involved in this community many years ago, as did you. And it is something where there is knowledge in that community that should be widespread, and I guess it’s been, I think it was Flagstaff, which was 2014, I believe.

[00:13:53] James: It was the next one after the 2010 one.

[00:13:56] David: Yeah, exactly. I think it was Flagstaff where I started trying to say, you know, we need to get these things to scale. And it’s something where that’s, you know, over a decade ago now, and I would argue, one of the failings of this community is that their knowledge is not widespread.

There’s such a depth of knowledge and expertise which understands in a particular classroom how to do wonderful statistical education, how to get people to enjoy the subject, how to get people to engage with data well, how to be thinking critically about this. But scaling that has been hard, and it’s not for lack of efforts. And the US has had this GAISE report, which was its effort to have a top down approach to really scale these efforts. New Zealand has inserted statistics, or data, in the curriculum from the first year of primary right the way through schooling.

So there are elements of scaling, but it’s been hard work and there aren’t the studies I can put to, the large scale studies, which are demonstrating this is an easy way to scale some of these good practices.

We are involved with work with Oxford on these randomised control trials for parenting initiatives where we now have publication about to come out, which would say, look, this is what we did, this is the effect it has, this is how you can do it. And that enables scaling. And within the statistical education community, there just haven’t been enough of those studies.

And that’s something which I have hoped that others would do, and I guess, part of the question is who. There’s a community of people who should be doing this, but my current feeling is it might be something that we need to push and be more involved in than I ever thought before.

[00:16:01] James: Could it be that it’s difficult to scale?

[00:16:05] David: No, everything’s difficult to scale, so that’s not the reason. Just ’cause it’s difficult to scale isn’t a reason not to do these studies. I think it’s more that, well, these studies are expensive. You need someone to lead a big EU grant or a big NSF grant, or go to a foundation and get a big grant to get a big collaboration of people showing how this can be scaled across context in different ways, doing those studies, showing that it’s impactful.

And I guess, within the community, most of the statistical education research has been focused on classroom level studies. There’s not been the ambition…

[00:16:52] James: Sort of like a big study to consider outside classroom across organisations.

[00:16:58] David: There have been big EU collaborations, for example, there’s a wonderful one on pro civic stats, where they’re trying to get the official statistics into schools. And I’m not saying that’s a bad idea, but I’m saying we actually need those studies based on what is known, not on something new, we need to get the existing methods, which are known to be working out at scale. And so we need work to happen on what does this look like in terms of the tools, in terms of the resources, in terms of actually getting this in the curriculums or working with curriculums so that these things can be scaled and kept at scale.

There’s been the International Statistics Literacy Project, which has been, in its own right, extremely valuable and powerful.

[00:17:49] James: Yeah.

[00:17:49] David: But it could be more so if it actually engaged with, well, how do we get statistical literacy more prominent in schools? How do we bring this into your geography class, your history class, so that you actually have the essence of learning with data embedded in student experience, not necessarily changing curriculum, but getting what is known to be effective.

It is known that if you teach with data in something like geography or history, you can get people not only more engaged in the subject itself, history or geography, but also in the use of data and how to be careful and how to actually learn scientifically. These are so important in our times.

[00:18:37] James: That sort of reminds me of, I think there was a talk at the ISI or there were discussions about… to get these things to scale in different contexts. You know, contexts are also very different, they vary. What works in Africa may not be what works in Europe.

But you mentioned geography, history, you know, including data there, there was a big discussion recently that one of the big gaps is actually equipping the teachers with the skills, they don’t have the skills and so that makes it even more difficult to implement such initiatives at those levels.

[00:19:12] David: Absolutely, and this is where, I’m not saying this is easy, but I’m saying it needs to happen, and it can happen, and it should happen. It needs people who actually have the depth of understanding from the statistical education community to engage in the challenge of really scaling these things.

[00:19:32] James: Yeah.

[00:19:33] David: I guess there are efforts within ISI, the International Statistics Institute, the ISI Academy is a new initiative where they’re trying to say we need to take this network that we have, this incredible resource, which is global volunteers passionate about statistics and data, and we need to channel that into a world reference center, this academy, where others can come and get quality.

A big part of that, and one of the pushes that I would be arguing for is that we need that to not just focus on official statistics, but to really lean into how do we make sure that the statistical training in general is widespread and effective at demonstrating the power and the limitations of data.

[00:20:35] James: That’s interesting. Yeah, I heard about the Academy during the Congress. It’s good to hear that one of its mandates would be to push these initiatives, these ideas.

[00:20:45] David: I’ve been fortunate enough to be part of the ISI Council. I’ve just stepped down from that responsibility last Friday. But one of the things that I did put a bit of effort into is supporting the launch of this ISI Academy, and it is something which I believe, you know, I’m very much on the periphery of now, but I do believe that done right. This could fill a gap, this could really be present in a way that taps into this community that does know how to do things and help be a source of reference as part of a collaboration to do this. ‘Cause they’re not gonna do it on their own, they are gonna have to do it in collaboration.

[00:21:30] James: Yeah, exactly.

[00:21:31] David: One of the challenges I think for me over the next few months is to see, can we bring the ISI Academy into some of our work and actually have this as something which helps them to build and to grow that reputation, which I think will come quite naturally because of the community behind it.

[00:21:50] James: I’m assuming the ISI Academy will have representatives from all the, you know, associations of the ISI.

[00:21:57] David: Not really at the moment. This is the thing, it’s sort of very small, it needs to get started, it needs some focus to get started. Because of the people who have been recruited to lead it, its focus has been suggested to be official statistics working with the national stats offices and so on.

And I’m not saying that’s bad, but my voice on this as part of the ISI capacity building team, which is involved in setting up the ISI Academy and so on, was to say yes, but don’t forget the education, work with IASE. Now this is the one association which can set you up for success.

Building that relationship between these associations, because these are volunteer associations, the ISI Academy is aiming to be more, but you know, IASE, ISI, these are volunteer associations. And so there can be challenges in terms of actually building that collaboration because people are often at their limit in terms of what they can do for their own association. And so looking to contribute to others can sometimes be more than they can contribute.

So it’s challenging, but I do think the importance of these, particularly IASE, engaging in the ISI Academy and enabling it to serve the education statistic and data education communities by providing the skills to do so. And, similarly, the ISI Academy, then elevating the work that’s done by IASE so that it can start to scale, maybe it’s the natural lead on these projects that could be these big ambitious projects for statistical education and the scaling of it.

[00:23:50] James: Yeah.

[00:23:51] David: This is what would be an obvious approach to me if we could actually navigate that collaboration.

[00:23:59] James: Yeah, you mentioned the National Statistics Services and I’m reminded of how big they were at the ISI, and an interesting aspect of what came out. I think some of them attended our sessions on statistics education, and there seems to be that synergy already that, you know, we need to be establishing these collaborations, because the NSOs, they have data and making it possible for that data to be used in the education could be something useful.

[00:24:25] David: Well more than this. In the African context, and this was one of the things that I learned really strongly from the conference in Zambia, which was an ISI sort of conference held in Zambia. This, you know, in the African context, the importance of academic statisticians and national statisticians working together is even more so because the communities are too small.

In some countries, like the UK, the community of official statisticians is big enough to be a community in its own right, and there’s plenty of academic statisticians doing their own thing, and they have their own communities. So that enables, in some sense, a level of separation and specialisation, which has advantages and disadvantages.

In many African contexts, they cannot afford that level of specialisation. Your statisticians go backwards and forwards ’cause there’s not enough statisticians in the NSOs and there’s not enough statisticians in academia. And so you actually often find somebody through their career will have times in both.

And so you have more people who bridge from one to the other, or who could bridge from one to the other. And the challenge is to actually get these communities to align so that they’re pulling in the same direction. These are interesting challenges, they’re very valuable, and there is an opportunity, as we’ve often talked about, for, in certain respects, the African contexts to lead to opportunities to lead, to really create these opportunities to lead.

[00:26:03] James: Interesting.

[00:26:05] David: When you said there were people in our sessions, they were mostly from African contexts.

[00:26:09] James: Yeah, they were mostly from Africa, and they were pushing for, you know, for us to explore these collaborations.

[00:26:15] David: Exactly, because they know in their context, and I remember particularly Tanzania, wonderful, wonderful lady from Tanzania, who was the former head of the national stats office for 17 years, has just stepped down, but is a powerhouse. She was the person who was driving, you know, and saying, look, we need this, what you are doing with the universities, this needs to also come to the national stats offices.

[00:26:39] James: Yeah, exactly.

[00:26:41] David: And I think she’s absolutely right. That collaboration just has to be there. I believe, James, that she’s got her eye on you now, you are in trouble.

[00:26:48] James: Yeah, yeah.

[00:26:52] David: You’re gonna have to make it happen.

[00:26:53] James: Yeah. gonna have to follow up with her.

[00:26:55] David: She’ll pull you in and you will support them to get this collaboration.

[00:26:59] James: Yeah, yeah, yeah.

I think it was an interesting event in Netherlands.

[00:27:05] David: Yeah, it was very nice, it was really interesting, inspiring. Ironically, of course, one of the most inspiring things for me was meeting an old colleague who we’ve known each other for forever or known of each other forever, but never really got down and discussed. And we had a, a short meeting together, he sort of reached out to me on the app and said, oh, we should meet and just catch up.

So we met in one of the coffee breaks and then we said, no, that wasn’t long enough, let’s have dinner. And so we then had dinner, and we actually spent a number of hours just talking about different experiences, you know, different areas, and I’m pretty sure we’re gonna work together in the future.

What we found is that our alignment is, you know, we’ve been working in parallel for so long, coming to a lot of the same sort of challenges, recognising the needs and the opportunities. And so hopefully, and this is what these events are all about. That will lead to collaborations of people who know each other already, but in our day-to-day work, we don’t have time to sit down, talk things through, and imagine what else could be. And one of my highlights of the event was that.

But I should say another highlight, and it does show that I’m getting old and I’m sure you’re gonna there pretty soon. When I first started going to these events, I wanted to go to the big talks to see what I could learn and what I could take. And a highlight for me was actually going to some of the talks of people who didn’t have much of an audience. Giving them an audience of someone who can listen, who can understand what they’re doing, and then who can comment in constructive ways, and maybe push them forward and challenge them to do more.

And I had people who in the audience when I was talking, when I was starting out, who I really appreciate and took a lot from. And now, to be able to repay that as well and to be in the audience and to see these exciting initiatives coming out of youngsters, for whom it was their first event.

[00:29:12] James: Yeah.

[00:29:13] David: And being able to sort of encourage them, say, yes, the work you are doing is good, is valuable, keep going, and have you considered this? That was a real privilege as well to now find myself in that position.

[00:29:26] James: Interesting. Yeah. I think for me, I think the highlight was to reconnect back to the community. I think it had been a while since I had been part of these conferences. And yeah, so I think I was able to meet people that I’ve known for a while, but I’ve not interacted with them. And I think the other thing is, I think I have made new connections with people in Africa, as well as even in Finland who were in the workshop and I had not met them in Finland.

For me, I think they’re going to be exciting sort of collaborations, you know, coming out of attending the ISI. So, let’s see.

[00:30:06] David: It’s ironic that you’ve had somebody in Finland who is based a few kilometers from where you are now based.

[00:30:13] James: Yeah.

[00:30:14] David: For the last year and a half that you’ve been in Finland, you’ve never met.

[00:30:18] James: Never met, yes.

[00:30:19] David: You meet at the World Statistics Congress and now suddenly you say, oh, we’re just down the road from each other, we should meet, and now you’re gonna meet.

[00:30:28] James: Yes, exactly.

Yeah. So I think that was very interesting.

[00:30:33] David: Anyway.

[00:30:35] James: The next one is when, ISI?

[00:30:37] David: The next ISI is two years from now, well, it’s a little bit less than two years because it’s earlier in the year.

[00:30:42] James: In Korea.

[00:30:43] David: That’s right. Let’s see whether we can make it there too. There’s ICOTS first next year.

[00:30:49] James: Yeah.

[00:30:50] David: Hopefully we’ll reconnect with the community in Brisbane, Australia.

[00:30:53] James: Exactly.

[00:30:56] David: Thanks.

[00:30:57] James: Thank you, David.