Dr Rebecca Goldin is a Professor of Mathematics at George Mason University and the Director of STATS, a collaboration between Sense About Science USA and the American Statistical Association (ASA) to improve statistical literacy among journalists, academic journal editors, and researchers.
Her work with STATS has appeared in many media sources, including NBC, CBS, NPR, CNN, and the Washington Post. She has worked individually with journalists from the Wall Street Journal, New York Times, 538 and many other media outlets, as well as run workshops for journalists and students alike.
She continues to work with journalists to improve statistical reporting and gave the Campion Lecture during last year’s Royal Statistical Society Conference.
Alison Oliver talks to Dr Goldin about her career so far.
1. Please could you tell us about your educational background and what was it that inspired you to pursue a career in mathematics and statistics?
My work in statistics began very late, but my work in mathematics began very early. I did a lot of math when I was a kid, but I didn’t think that I liked it at all; in fact, it was things that I thought were not math that excited me, like puzzles and number patterns. I became really excited about ideas that were not part of traditional math. Like for many people, there were concepts that tripped me up, but then you move forward. I got quite accelerated in high school in mathematics partly due to studying math on my own. That gave me this false belief that I really knew what math was, and I went to college actually thinking math wass the one thing I know I’m not going to do. But then I took some advanced courses and soon enough ended up taking a course in topology. The course introduced me to very beautiful mathematics, and I fell in love. So I ended up staying with mathematics.
2. Your research interests include symplectic geometry, group actions and related combinatorics. Among your research interests are: Schubert polynomials and the intersection properties of Schubert varieties, toric varieties, equivariant cohomology, moment maps and symplectic and hyperkohler reduction, and orbifold cohomology. What are you working on currently?
Recently I’ve been working on intersections of some very strange spaces, called flag varieties, that consist of sequences of embedded vector spaces, one inside of another. People ask questions about how some specific subvarieties intersect. I’ve been trying to describe these intersections using some geometric invariants, which are fascinating in part because they take into account the fact that we have groups that are actually moving these spaces around. The groups are symmetries, and you try to exploit the symmetries to explain how the vector spaces are intersecting. Right now I am trying to derive combinatorial formulas for the equivariant K-theory of flag varieties.
3. How did you become involved in the media’s use of statistics?
It was a strange set of circumstances by which an organization, STATS.org, came to George Mason saying, “We’d like to be affiliated with the university.” The dean at that time was very excited about this opportunity, and he asked if I would be interested in becoming part of their efforts, and I became the Director of Science. At that time, the effort was to try to correct misstatements put out by researchers, or by government organizations using—or abusing—statistics. At that time, we called it the Statistical Assessment Service, and we use STATS as kind of a fake acronym for it STATS.org was a “gotcha” kind of organization, where we’d point out flaws and errors that were being picked up by the media. We realized that actually it’s much better to work with media rather than criticize after the fact. Over time, the organization changed its funding structure as well as its institutional structure and goals. By 2014, STATS.org was no longer, but STATS was born as a project of Sense About Science USA. That’s when I became the Director of STATS. We now have a much more educational focus than we used to have. We try to talk to journalists before they’re actually publishing things, before they’re putting out something, that may have a mistake in it. We try to promote a certain understanding and allow people to write from a perspective of knowledge rather than just trying to stab at the words and not be too descriptive because one may not be entirely sure what is going on. We allow journalists to really have access to statistical knowledge and to be able to use statistical and mathematical reasoning to make their stories better.
4. What are the main challenges that STATS has faced so far?
I would put them in two categories. One challenge is that it’s really hard work to get in touch with journalists, to be talking to them all the time, to have institutional structures to support them. Journalists don’t get rewarded for reaching out to statisticians to get support with their work; they get rewarded in terms of some of the quality of their work, but they don’t receive explicit rewards, so there’s not a structure within their news organizations for them to do it. Often they are under a deadline, which discourages them to seek professional advice about the mathematical reasoning they use. Similarly, it’s actually really hard to get our workshops into journalism schools or newsrooms. We’d like to get support from the journalists or editors themselves, to recognize that this is something that is really huge and can really promote their work.
A separate very real challenge is the funding for STATS, because we don’t want to accept money from industry. There’s a lot of interest from industry to support work that may be biased in favour of industry and we can’t accept it. We’re very concerned about losing our reputation with journalists who may see our aims as nefarious if we were to accept industry money. We have a policy never accept money from industry; that makes it very challenging just to get the money to support our efforts.
5. What kind of feedback are you receiving from the media community you have worked with so far?
Most of the people that we’ve worked with have been extremely grateful—so thrilled that there’s a resource that they can ask the stupidest question or the smartest question – they can ask whatever they want. They feel really comfortable. But they also feel like there are very few places for them to obtain that knowledge, so there’s a lot of gratitude just for our existence and our ability to be there for them, for the quickness with which we respond to their queries. A lot of very positive feedback from media.
One of the things we are also hearing from media is that they wish they had a lot more access, support, and knowledge. We give them surveys about what it is they want to learn about statistics, and they usually list there what would be involved in a full-year course in statistics. Of course, the way you would teach it to a journalist isn’t the way you need to teach it to someone who might practice statistics, but I think that they typically list maybe ten topics that they really want to get their heads around because they’re reading research papers that use statistics. They’re really involved in things that require statistics, and they want to understand what the results are. Sometimes they are themselves interested in data science or data journalism, and they want to know how to actually do some analyses. So if anything, a big component of the feedback that we get, especially from our workshops, is ‘can you come back’? ‘Can you do more?’ ‘Can we do something longer?’ ‘Could we fit in more topics into a small period of time?’ I think there’s also sometimes a misunderstanding about how much information you can get across in a two-hour workshop. For example, if we do a workshop, we might only talk about three different topics in depth, and I think they feel like, ‘oh, couldn’t we have gotten to ten?’ If you ask them beforehand what it is you want to cover in those two hours, they’ll give you ten topics that they’re really excited about. Of course, they also want examples and case studies, and it’s just impossible to do it in such a short period of time. But if anything, I think what we hear is ‘more please.’
One of the things that we are trying to do, and we’ve applied for some funding to do, is to create some online kind of app-type modules that allow people to engage with the ideas in a tutorial fashion but with an active involvement of the user. We hope to design them in a really beautiful way so people want to be there —kind of like some of Hans Rosling’s work, where your instant reaction is “It’s just so gorgeous.” It’s very compelling! So, to try to engage people with that kind of fine design mixed with really good ideas explained while somebody is engaging actively with an interactive app that gives them feedback as to what they’re doing—that’s something that we’d really like to do, and we’ve been thinking a lot about how we would design it and what are the topics that we’d include. It could then be implemented into other people’s coursework, if they’re teaching a course in journalism; it could be used individually by people who are just curious and want to try it; it could be put into some kind of case study environment or embedded into a lot of different kinds of media.
6. Your lecture at RSS 2017 was the Campion Lecture entitled ‘The Media’s Love-Hate Relationship with Statistics: Challenges in Communication’. Please could you tell us more about this topic? What was the one thing that you wanted your audience to take away from your lecture?
I think the one takeaway was that journalists really want to do right, and they really want to understand things. They have very serious challenges, but they also have some strengths that are not strengths of the statistical community. There should be a lot of respect for what the profession of journalism is, what the fairly unique challenges that journalists face within their work are, and what their fairly unique strengths are. Journalists want to do right by us. So if there’s a certain amount of patience and the right kind of angling to talk about statistical topics, journalists really can engage with those ideas and “get it.” Notwithstanding of course that there are times when a statistician is not going to be successful at explaining something, but trying to communicate is a very worthy endeavour because the impact is so much greater than the individual who understands it. That’s the pathway for communication between scientists and the public, so it’s probably the most important kind of individual person you can convince to think quantitatively.
7. What do you think have been the most important recent developments in the field and will these influence your teaching in future years?
To me, the biggest changes that are coming about now have to do with data journalism, the availability of data and the social interest in data. I think that is really huge—it’s impacting the statistical community in terms of data science and the development of really serious programs in data science. Data science is not just a branch of statistics with the involvement of mathematics and statistics and computer science , but rather a field in its own right. It’s also really impacting journalism, since data is so readily available. It’s like a big huge train that’s arriving and/or maybe going to hit us.
But does that impact how it is that I teach?. I think it does impact our work because there have been a many queries and conversations with journalists about how to analyze data they have collected or accessed. At times they ask questions like, “how would I take this Excel chart with all this information on it and try to turn it into something useful?” or “How do I actually make a compelling graph that tells my story?” or “How do I do all of these things that I think are on the surface of what we call data journalism?” At the same time, I think the newsroom format will change over time—the writers in the room will evolve. Media companies are going to have to hire people who are specialists in data science or who are statisticians. I heard that recently the BBC hired an RSS fellow from the statistics section. That’s really great. The involvement of people who are professionally trained in this kind of work into the newsroom in a direct way is fantastic; I wish it happened more.
There’s also probably greater need for support in data journalism than is being hired. The need is going to grow before it’s met, if ever the need will be met to the level of what would satisfy those of us who work in quantitative sciences. It may not happen at Fox News any time soon. The starting point would be the Wall Street Journal and the New York Times. A lot of these organizations are reluctant to invest in really serious ways; instead they invest in people who are trained as journalists who have a strong quantitative background. They are not taking people who are trained as statisticians who have a strong background in journalism. It’s a different inversion of the skillset that I really feel should be central. There’s also really serious financial constraints to their hiring experts, because statisticians are expensive and it’s not easy to find good ones who can communicate well with the public. Those really good ones are so good that they get other jobs. So I think it’s a real challenge for them to find the right people, but it would help if they really wanted to have that there. I think that’s going to happen to some extent, but probably not as much as we’d like it to happen.
8. Your research has been published in many journals and books: is there a particular article that you are most proud of?
I don’t know, I spend a lot of time thinking about things that don’t go into papers, as well. I think that from the point of view of productivity, sometimes I am too busy to get it all written down, and that’s not a good thing.
Part of it is also that I do work in such different different fields. I was very excited about some work that I had written for STATS that was about criticizing a couple of scientific studies on chronic fatigue syndrome, and I felt I wrote some really excellent stuff there but it wasn’t really what you’d call traditional research work.
I’m working on a project in mathematics now that is very very exciting to me about something called equivariant K-theory, and that is about some geometric invariance of these spaces of embedded vector spaces and how they intersect; it’s just really beautiful mathematics. I have some really nice results that I’m in the midst of writing up, so I’m very excited about that work.
Copyright: Image copyright of Dr Rebecca Goldin