Louise Marie Ryan is an Australian biostatistician, a distinguished professor of statistics in the School of Mathematical Sciences at the University of Technology Sydney, She is known for her work applying statistics to cancer and risk assessment in environmental health. Louise is also the Editor in Chief of Statistics in Medicine. She has received a number of prestigious awards and honors, most recently her 2012 election to the Australian Academy of Science, a 2015 honorary doctorate from Ghent University, Harvard’s 2015 Centennial Medal and her new role as President of the International Biometrics Society.
+ Full Transcript
Rosemary Pennington: Uncertainty in life is inescapable no matter how much we wish it were otherwise. In the realm of science and statistics uncertainty is the rule. Communicating what uncertainty means in terms of research is an, at times, herculean task, it’s also the focus of this episode of Stats and Stories, where we explore the statistics behind the stories and the stories behind the statistics. I’m Rosemary Pennington. Stats and Stories is a production of Miami University’s Departments of Statistics, and Media, Journalism and Film, as well as the American Statistical Association. Joining me in the studio are regular panelists John Bailer, Chair of Miami’s Statistics Department and Richard Campbell of Media Journalism and Film. Our guest today is Louise Ryan. Ryan is a Professor of Statistics in the School of Mathematical Sciences at the University of Technology Sydney. President of International Biometric Society, and the Editor-in-chief of Statistics in Medicine. She’s also written about the issue of uncertainty. Louise, thank you so much for being here.
Louise Ryan: My pleasure. Thank you for having me.
Pennington: How did you find yourself writing about uncertainty?
Ryan: Well, it’s a long story.
[Laughter]
John Bailer: We have time, we have time, Louise.
Ryan: Well, you know as a kid I always was fascinated by mathematics and always found that the world was a little bit chaotic and hard to understand. I found that math was the way to try to find patterns and help to make sense of this chaotic world. When I went to University and discovered statistics and probability it really started to click, and I suppose through my career I always liked working on problems that connected the mathematical and statistical sciences with the real world. And I just gradually, was drawn to those kinds of problems and got involved with a few projects where it was really trying to tackle those really difficult questions about when, say, a government agency needs to make a decision about an environmental exposure or environmental standards, there’s lots of uncertainty, and statistics can help make your way through that uncertainty to help come to a decision. So that really appealed to me.
Bailer: So, what are you working on now that’s forcing you to deal with this decision making under uncertainty?
Ryan: Well yes, I’ve got a few projects going at the moment that had that play. But I’ve been working with New South Wales Office of Chief Scientist and Engineer, and they operate a little bit like the Academy of Sciences in the US, where they’ll often be approached by government agencies faced with a difficult decision to try bring together some experts to help advise. So, I’ve been doing some projects with them and they get all the difficult projects. The easy ones are not needed. But the difficult ones are things like how government agencies should be making decisions about water usage and water allocation in Australia. Australia is a continent where our climate is quite extreme and we’re quite prone to droughts. So, managing water resources is really challenging. So, I’ve been involved in projects where my task has been to try to bring some of that quantitative thinking to help make decisions about how sustainable our water resources are. Does it make sense to allow access to these deep-water aquifers and that kind of question. It’s a topic that’s in the news a lot in Australia these days.
Richard Campbell: You mention the news there, so when you’re reading the news and you’re a statistician, you understand uncertainty, a lot of times the news often gives us numbers and statistics that make us seem certain. They don’t talk about things like margin of error very well, or certainly uncertainty. For audiences that don’t really understand statistics. And you might also talk about some of the criticism you might have of journalism that’s not reporting some of the complexities in the work that you do.
Ryan: That’s a tough question. I think that I’ve seen some good expository writing lately, some journalist writing Popular Science, where they are talking about the fact that unlike what a lot of people think, science is not just a hard, cut and dry topic. Science is something that is constantly evolving as we improve and gain new understandings and so on, so trying to communicate to the public that science is actually an evolutionary process of this huge community of scientists across the world trying to make sense of the reality of Universally living, that it’s an evolutionary and changing process; it’s not static. So, getting that concept across, I think is very helpful. And I’ve definitely seen a lot more people starting to write in that way, in recent years, and I think things like climate change help with that. People are getting a clearer sense that it’s not really just black and white. It’s quite subtle and there is uncertainty. So, I think the public is becoming a little bit more aware of those kinds of issues. But in terms of the concrete elements of how you communicate these things, I think that it’s a mistake just to write in a very dry, technical way with lots of facts and figures and uninterpretable graphs. But finding ways to present information in an interesting and engaging way with interesting graphics and so on, you often start to see, these days, more and more people trying to find ways to present information in a way that is visually appealing. You often see things where you see graphical constructions showing how things change over time. You might have seen some of those on Facebook, and those can be really interesting. Even people who are nonmathematically oriented really like those things because you can see how things are changing. So, finding interesting ways to represent complex data, and particularly changes over time, I think can be very helpful. And I think the classic statistical training that we give our students in University doesn’t really prepare them for that, because it brings in a lot of artistic elements, and communication elements that are critically important, and you need people who are skilled in those areas. I’m really impressed with some of the stuff that’s coming out around data journalism and particularly some of the graphical things people can do. So, some of the things that we’ve done with our Chief Scientist Office work is, for example, constructing interactive graphical tools that potentially the public can go and play with and say how did you get this border change look over time, or how did this aqueduct look over time. Letting people interact with the data is really helpful.
Bailer: The visually appealing and we certainly agree with that a lot, that’s a great point. But do you have other strategies that you routinely employ to communicate complex information?
Ryan: Well, I don’t know that I’m as good at it as I would like to be. I try to find ways to write so that you don’t over emphasize the technical side of things. You try to avoid having big sets of numbers and tables in favor of graphs wherever you can. [] has written a lot about this thing, well lots of people have. The way you present information can have a big impact on people, so for example if you’re trying to encourage people to consider going into a clinical trial for- maybe they’ve been diagnosed with cancer and they need to go into a trial- you can present it in two ways. You can say 5% of the people on this trial getting this treatment likely to die or you can say 95% are likely to survive, or you can do it in a more balanced way and say 5% might die and 95% might live. It’s how you frame it; a lot of work is being done on framing communication. Understanding how people respond to the information is really an important topic. It’s not something that I feel like I’m an expert in, but I’m very aware of, it and try to keep that in mind as I’m communicating these ideas.
Bailer: So, I’m curious, you mentioned water usage and allocation, could you help unpack that a little bit in terms of some of these uncertainties that are part of this water allocation/water usage problem?
Ryan: Well it’s quite a complicated issue. Over the last few months I’ve had the opportunity to read a bit about hydrogeology literature, where if you really want to know- suppose you’re in a community that needs to access water from underneath aquifers, and increasingly that’s the case. I’m sure you might have heard of a town called Alice Springs, which is in Central Australia, that large iconic rock that people recognize as being the middle of Australia, well Alice Springs which is the nearest big town, all of their water comes from an underground aquifer. And it’s interesting I was there just a few months ago and I was reading about the water, because I was curious like where does the water come from? And it turns out that it comes from the aquifer, but they don’t really know how sustainable it is, where it comes from, where the origin of the water is, how long it’s been there. So, to really understand the nature of these underground aquifers requires huge amounts of investment, of geological studies to understand the nature of the underground rock structures, the characteristics of the rock, the plurality of them and so on, how the water flows through, where the water flows from- so getting all that information is extremely expensive. Sometimes if you’re in areas there’s a lot of mining, they’ll have done those explorations because they want to know what’s down there in terms of minerals, but the same kind of info is needed to try to understand the nature of the structure, and what’s underneath the ground, so you can understand how the water flows through, where it goes to, and where it sits, and so on. So, as you can imagine there’s massive amounts of uncertainty in understanding those structures. Once they build the models which are very complicated, computer-based models which can take days to run, there’s uncertainty about the parameters that go in to describe those models and so on. In practice, when decisions are made about water allocation it will often be fairly simplistic things such as, well, we assume that we can safely take 6% of the water recharge each year. Which may be 6% of the annual rainfall; that amount can be safely extracted from underground aquifers. So, there’s a lot of uncertainty in Australia in particular, I’m not sure about the US, it gets extremely political and there’s a lot of money connected. Recently there was a big controversy where one of our politicians got in trouble because he was connected with a company that had actually bought up property purely because they wanted the water rights. And then when there was a drought and there was a lot of concern about water access, they sold these water rights for a ridiculous amount of money and of course it hit all the headlines. So, this is where that uncertainty translates to money and politics and so on. So, it can easily get very- everybody can get worked up and it becomes a highly emotional issue. Yeah, I could go on for a long time.
Pennington: You’re listening to Stats and Stories and today we’re talking to Louise Ryan of the University of Technology Sydney. So, you were just talking about the political complications around this one particular issue, I wonder, it seems like we’re in a moment where we have a better understanding of where the uncertainty lies in work. But at the same time, when people feel less inclined to trust research. So I guess I wonder how you think about navigating this space where it feels, again, on the one hand, that we have a better handle on what the uncertainties may or may not be, and yet people who will look at any uncertainty as a reason to discard the research and how you think about navigating that space.
Ryan: Yes. People can do that, because everybody’s got their own agenda. People have their own agenda for different sorts of reasons, so lots of times, and I’ve seen this happen quite often. people will tap on to a certain source of uncertainty and say, oh well we don’t know the sustainability of this aquifer therefore we can’t take any action whatsoever. So, people can use it as an excuse to not act. And it can be put up as a roadblock, so you have to be really careful about that kind of thing. In my opinion, endless talking and endless arguing doesn’t get you very far. I’m a big believer in providing data; giving people data that they can use to see what’s happening over time; monitoring data. I’m sure you’ve heard of Australia’s Great Barrier Reef, which is one of our National treasures, a big source of beauty and an attraction of Australia, but also big source of tourist dollars and so on, and there’s lots of talk these days about how the reef is under stress because of climate change, and coral bleaching and things like this. But one of the things that you can do is put systems in place that monitor how the reef is doing. And if you have good technical resources in place where people can see those data, and how they are changing at the time, and access those data in a way that doesn’t require them to be experts- so easy to access monitoring data that lets the public get a hands-on feel, I think that can make a big difference, and that’s one if the things were pushing for in this water work is making sure that people have access to good quality monitoring data so that they can see what’s happening. And if people are worried, for example, that the water resources are dwindling accessing the monitoring data and saying, well, they’re actually okay they are not dwindling. So actually seeing some evidence I think can make a big difference, but that’s where some of these tools, these interactive data analysis or interactive graphical tools can make a gigantic difference, but they have to be user friendly and they have to be attractive in my opinion. So that’s where that interface between statistics, art, and computer science comes into play.
Campbell: You know I read your piece in the conversation and across the top of that piece they talk about the fight to defend research evidence and facts, and you’re talking a little bit about that, and you’re kind of commitment to doing a better job communicating to the public is very interesting. I first want to know where that comes from? And I’m also curious about whether the same thing we deal with here in the states- where people are suspicious of facts and data and evidence, is this something that you face also in Australia?
Ryan: Absolutely, we face it in Australia. I think its everywhere these days, and it’s complicated. Scientists complicate it. Think of the climate change debates, and I don’t know if you’re a Facebook user but you get on to Facebook and you see all sorts of ridiculous posts where people are coming down one side of it or the other, but that’s why I always think come back to facts, show people some facts, and show people some data, and let them engage with the data themselves. That, I think, can make a big difference. People love data, most people really do. Sometimes I see my sister respond to some interesting graphical display, and people love it, but they love it when it is done in some dot of interesting way that us visually attractive and appealing as well as communicative. So, there’s just so much potential there to do more than that. It’s starting to be done and I’m not saying I can do it, I don’t have those artistic abilities, but you need that fusion with statistics and art and computer/web savvy to be able to pull those things together. But I think in terms of how I’ve gotten interested in the topic, it’s through involvement with Chief Scientist Office or when I lived in the US, I lived in the US for 30 years, and I was very privileged to sit on a couple panels for the National Academy of Scientists, where we were asked to tackle these really nutty questions, and I realized then, how politically loaded these questions are, and they only get to that level when it’s really hard to make a decision. So setting standards for us, say in drinking water for example, was one that I always remembered had a profound influence on me, because I was a statistician on the project but I got to think a lot about how people respond to these kinds of data and these kind of questions, and what can I do how can I use my statistical tools to help make sense of that? So, it was involvement in those kinds of projects that I realized how important this is. And communication is multifaceted. It’s communicating to your fellow scientists, to the policy makers, to the politicians, and it’s communicating to that public. And they all have different ideas about what is important, and what their priorities are.
Bailer: I like your observation and your thoughts about the use of data and visualization to inform the public and change attitudes. In essence, it sounds like you’re making data accessible. When you were talking about these very complicated water models, these hydrogeology models, one thing that I wonder about and I think a lot about is the idea of data reducing uncertainty. So, when you’re thinking about these hydrology models, so you think about data collection or particular variables that help reduce uncertainty with theses models?
Ryan: Yes. Again its something I’ve only been learning about relatively recently, so I by no means consider myself an expert on these hydrogeology models, but it seems like the real interesting stuff happening these days comes about when people build these process based models, so you have all the hydrological theories about how it flows the shape of the rocks and underground structures and so on, and you fuse that together with empirical data. And these days they’ll do lots of things, like they’ll put down these what they call monitoring balls. Basically it’s a deep pipe that goes down through the rock into the aquifer and they put a little recording device at the top which can measure in continuous time, basically, how far down they have to go before they find water which tells them something about the pressure in the underground aquifers, and how much water is there. What’s interesting is how those levels change over time. So you can collect a lot of that kind of data very easily or relatively easily, and if you have that rich set of monitoring data you can start to build these models that fuse together the statistical perspective with the physics and hydrologically based perspective to come up with these very rich, albeit very complicated, models of the system that you’re trying to understand. Sometimes you hear a little bit of a dichotomy, there’s the hydrologist versus the statisticians, and those are two completely different worlds. But in fact, there’s a lot of ground in between where those worlds can come together and build rich much more interpretable models. Because you need the empirical data to hone the underlying hydrological model.
Campbell: In one of your writings or maybe on another podcast I listened to, you’re reminding me of a phrase that you used that I liked a lot. I wanted you to talk about it, and it was about the virtuous cycle of collaboration. And you talked about collaboration being one of the things that motivates you instead of just being drawn to theory and working alone, that collaboration is something that I think when I hear you talk and read your writing you’re very enthusiastic about this, can you talk a little bit about that?
Ryan: Well yeah, that’s an interesting question and I do like to think about the cycle of collaboration some of my earlier training was of that more classic theoretical orientation, and you know, there are many statisticians who are of that mathematical orientation and what they love to do is develop theories, and prove theorems and so on. I’ve always been drawn to problems where I get to engage in a real-world question, but the thing is the real world is complicated as we discuss. So, if you want to engage in a real-world question, as a statistician I’m never going to have all the tools that I’m going to need to do that. So, if for example I’m trying to help the Chief Scientist Office, and there is a question about water resource allocation, well I need to be able to work with hydrogeologists and policy makers, I need to understand a little about how all the issues relevant to their field come to the table. It becomes this interplay between these different disciplines. And you need that in order to be able to solve the problem in a realistic way. And I mentioned before some of the really formative experiences I had when I was living in the states was working with the National Academy of Scientists there, and they have that down to a fine art. They knew that if they got a really complicated problem, like how we set standards for how much fish for women to eat when they are pregnant, because of methylmercury. The National Academy of Scientists in the US really does it well. They understand that to tackle complicated, politically loaded, economically loaded, complex real-world questions, you need an interdisciplinary team. And so that’s what they would do. They’d form committees, put together a group of 15-20 people representing all these different disciplines, and basically make them fight it out several months in a locked room. Of course, I am exaggerating a little bit. But at the beginning you think this is hopeless, we’re never going to find an answer. But gradually there’s this shared opinion that emerges. Sometimes there’s dissenting opinions, but basically you talk about the problem, you mull it over, you argue, you fight and eventually you come to some shared understanding and an answer emerges. Its just quite remarkable how it happens. There’s something quite magical about that process.
Bailer: And it’s a lot of fun.
Ryan: It’s a lot of fun yes it is. Very intense but yes, very rewarding. So, those experiences really influenced how I think about statistics these days.
Bailer: So, I wanted to ask you, you’re President of the International Biometric Society, what are you really excited about right now? Things that are going on at IBS?
Ryan: Well, lots of things actually. IBS is quite a unique society in that we are international, and we have 37 regions all over the world, and the global nature of our society is a thing that, I think, is most appealing about it, but it also means it’s a very complicated society. We started over 70 years ago, and the regions are all quite variable in terms of sizes and level of resource and sophistication and so on. So, it’s the communication aspect that s actually quite challenging in this society. So, it’s the communication that is quite challenging for this society. So, at the moment we’ve been working on revision to our IBS webpage, we are trying to utilize a community structures where we are trying to build facilities that allows people to communicate in a much more effective way, so again it’s all about communication. We are working with a company called Higher Logic, that specializes in creating these online communities’, formal technical communities like ours. So, we’re looking to, for example, to create a community for young statisticians, different regions might have their own communities, and so on. It’s in the early stages but trying to provide more opportunities for our members to really engage, tell us what they want, learn from each other, that’s what I’m really trying to create. And we’re also planning for our biannual conference, which will be in Seoul, Korea next year in July, so that’s taking up a lot of my time and energy for that.
Bailer: Well good luck.
Pennington: That’s all the time we have for this episode of Stats and Stories, Louise thank you so much for being here.
Ryan: Oh, my pleasure. Thank you so much for having me Rosemary, I really appreciate it.
Pennington: Stats and Stories is a partnership between Miami University’s Departments of Statistics and Media, Journalism and Film, and the American Statistical Association. You can follow us on Twitter, Apple podcasts or other places you find podcasts. If you’d like to share your thoughts on the program send your email to statsandstories@miamioh.edu or check us out at statsandstories.net, and be sure to listen for future editions of Stats and Stories, where we explore the statistics behind the stories and the stories behind the statistics.