Robert Cuffe is the head of statistics for BBC news. Before that he worked on HIV drug trials at GlaxoSmithKline and is head of statistics at ViiV Health Care. Cuffe is a statistical ambassador for the Royal Statistical Society and was chairman of the UK pharmaceutical statistician’s industry body PSI where he worked with the Science Media Center to set up a briefing service for the Lay Science Press. His research interests deal primarily with health statistics and the general communication of statistics as a whole.
Episode Description
Being able to effectively communicate data is becoming an increasingly important part of a journalists job, so much so that news outlets are expanding their staffs to include data scientists and statisticians and that is the focus of this episode of Stats and Stories with guest Robert Cuffe.
+Full Transcript
Rosemary Pennington: Being able to effectively communicate data is becoming an increasingly important part of a journalists job, so much so that news outlets are expanding their staffs to include data scientists and statisticians and that is 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 is regular panelist John Bailer, Chair of Miami’s Statistics Department. Richard Campbell of Media, Journalism and Film is away. Our guest today is Robert Cuffe. Cuffe is the head of statistics for BBC news. Before that he worked on HIV drug trials at GlaxoSmithKline and is head of statistics at ViiV Health Care. Cuffe is a statistical ambassador for the Royal Statistical Society and was chairman of the UK pharmaceutical statistician’s industry body PSI where he worked with the Science Media Center to set up a briefing service for the Lay Science Press. His research interests deal primarily with health statistics and the general communication of statistics as a whole, Robert, thank you so much for joining us today.
Cuffe: It’s a pleasure to be with you.
Pennington: Now, you studied math and, I believe, psychology and statistics in school, you worked as a statistician, how did you move from that realm into the world of news which, I would imagine, at times feels very different?
Cuffe: Very different, well as the bio suggests, while I was still in the pharmaceutical industry, I was doing small bits on the side, so working with the Royal Statistical Society to do stats outreach to the lay media and then working with the Science Media Center to set up that kind of service to help journalists get the stats in any – and then working with the Science Media Center to set up the briefing service for the science journalists. And so, what that was was to help deal with every medical story that comes out tends to have stats with it and an organization like the Science Media Center in the UK, which thinks of itself as a press office for science, really that helps brief journalists on the key parts of a story, put them in contact with great experts and they can put them in touch with people like David Spiegel for every medical story or every story that involves stats so we built for them a cohort of volunteers, of statisticians you know, five eight years of experience who would know enough to be able to critique most of the stuff you see in the medical literature but the type of profile that would put them in front of something like the Science Media Center. So, working with them we got a gang and PSI in the UK the former stats body worked on that, and between those two things I was kind of starting to get a kind of experience and then when the job came up in the BBC it was a terrifying step but a very exciting step to go from doing it in my spare time to for a living.
Bailer: So, what is a typical day for you like now?
Cuffe: At the moment it’s pretty chaotic
Bailer: Let me guess I bet COVID is part of it is that part of it?
Cuffe: Yeah.
Bailer: But in general.
Cuffe: Well, it really does depend. So, at the moment the job would be- let’s see, the average day we might have a- you know we might for example at half nine in the morning you might have a release from the office for National Statistics that would describe you know what the overall death rates look like what the excess mortality looks like, and so you might as soon as that comes out 9:30 the aim is to kind of get some copy out that the whole BBC can understand the main points of that story within about four to five minutes.
Pennington: Oh wow.
Bailer: Woah.
Cuffe: So, to get those the key bullets out and then after that maybe I’d have a little bit more time to kind of comb through the release in a bit more detail to put additional analyses out and at the same time I’d be working with a data scientist to pick the main charts that we’d use to describe that graphic online o the story that will describe it at the live page, at one of the ways that journalism seems to be moving is the use of kind of almost like a feed within a news organization like a twitter feed a live page where you put kind of short stories all the time so we might throw some graphics into that quickly. Working on checking that information is all correct and then you’d probably do a broadcast round on it around ten o’clock, so that would be on the news channel and that’s kind of you know three questions three answers summarizing the stats maybe with the graphics as well reversioned for TV and on a good day that’s the end of it and you can kind of get into helping other people do their stories later in the day but there will be frequently a couple of days where you do that, rinse and repeat a couple of times where there’s another release that comes out at 11:00 or 2:00 Fridays in the UK are just really busy where we have a big study describing how many infections there are in the UK and then the scientists advising the government they come out with the error rate across the UK and we’d also get kind of reports coming out from the public health body so theirs is just trying to cover the data here is like trying to take a glass of water from a fire hose. There are just nonstop numbers but the timing when you’re in that role of kind of facing the output that we’re putting stuff out for people the timings are pretty tight. And that’s very different to where you’re sitting in the back seat doing more advisory work or consultancy work helping people with their stuff.
Pennington: What was that transition like for you to go from, what I would assume is a much more- I would say maybe not relaxed if you’re giving drug trials, but a more leisurely compared to a news timeline sort of way of crunching numbers to suddenly being in this place where it sounds like within a half an hour you’re from press release to hitting you know the air and having to sort of talking about something intelligently. How was that transition for you? Like to go into a news environment where you were sort of expected to be able to make sense of these very very quickly.
Cuffe: It took me a while to learn it. I would say it’s very different. So, you know in drug development your deadlines are five years away. Trials take a long time to do and you know those deadlines they can be pretty painfully if you need to negotiate a shift with your management if a trial takes longer than you expected or something goes wrong. But deadlines in news are well, it’s not a deadline, you’re on, it doesn’t matter if you’re ready. You’re on. So that concentrates the mind, and it did take me a while to realize that. you know one of the first time I was helping out with the business desk with a slightly awkward set of economic stats I said to the desk editor on the day that’s great, thanks, okay, look just sit down I’ll read these for ten minutes and then I’ll tell you what is the- what the main points are and he said no.
[Laughter]
Cuffe: You know, the pressure for people to get stuff out quickly, I think, is something that it’s hard to grasp when you’re not in the industry. And I think some people would legitimately ask well why does it have to be like that? And it’s just gravity I don’t think there’s any point in complaining that it’s like that, it just is. And so, if you take an example of a one of the releases that we would get you know GDP you know are we going into a recession or something like that. and well if that’s a story and it’s a huge story, well then you know, in an organization like the BBC where you’ve got online, radio, podcasts, TV and you’ve got bulletins at 1:00 in the afternoon and 6:00, well you need to get the trucks rolling if it’s a big story. You know, you need to interview the chancellor and the leader of the opposition you know, one person to say well this is thanks to the glorious policies we’ve put in place over the last five years and the other one to say well this would have been much better if it had been under us. You need to get your word out in the street to say you know GDP doesn’t mean much to me it’s still- I’m still having a tough time, you need to get all that together, you need to pull the package together and you need to get it done in about three and a half hours. And of course, if it’s not a story well then you need to get something else to fill the half hour because no matter what you’re going to be on for a half hour at 1:00. So you need to make the decisions quickly and it’s not just in our setting you know in print media, you know you’ve got to make decisions fast because the newspapers are really the only barrier between the readers and everything that happened in the world yesterday and someone has got to make a decision about what goes in and what doesn’t go in and so the volume of stuff that the editors or the people who are putting the paper together are dealing with every day is just it’s mind blowing and do they have to be able to make decisions very quickly on the basis of very very clear information I mean that’s what’s at a premium.
Bailer: It sounds like a tremendous challenge, I mean I’m hearing you talk about the idea of there’s this issue of clarity there’s an accuracy and then there’s a timeliness. Is there some saying that you can have two of the three, but you can’t have all three? You know that-
Cuffe: Oh, I challenge that
Bailer: Good, good.
Cuffe: I think the skill of the journalists that I work with is their ability to be remarkably clear on very short order. And to do it- not just to do it on a time budget but also to do it on a budget of words. You know there’s only so many characters you can fit in a tweet or a headline or in, you know, the first four paragraphs of a story which is you know where you kind of get a- we aim to get the bulk of the story or the important things you need to take away in that short amount of information and their ability to turn that out quickly is I think really impressive and you’re not just doing- you’re doing it slightly differently when you’re doing it with numbers but every story is caveat, every story has nuance every story has complexity. Brexit is difficult, Coronavirus is difficult, not just for the numbers and journalists do that all the time. Numbers are just one part of the world that has to be summarized.
Pennington: I’m looking at the key points on the BBCs editorial guidelines for reporting statistics and there’s something that sort of sticks out to me because I teach journalism. I’m a former journalist. And so I’m constantly sort of pushing to my students, you know, don’t take anything a source tells you at face value, verification is the essence of what we do, you know it’s great if you use a quote but make sure there’s something of news value in the quote, and so one of the points you have is you know we ask for skepticism about facts and quotes but I’m just going to quote this avoid taking statistics at face value. Why is that the first point and why does that seem to be something that you’re trying to impress upon- I’m assuming the journalists who are using this as they’re reporting?
Cuffe: Well, every number has a history. Every number has a story that’s come from people and who’ve collected a particular set of data and you would need to ask the same questions of any number that you would ask of any other fact. So, who’s told me this why have they told me this is it reasonably likely to be true and what’s the basis for the assertion that’s being put in front of me? And things feel real with numbers. Like numbers seem solid and dependable and reliable and they can be but they’re just as open to any of the sources of uncertainty as any other thing that we hear about during the day and it’s just reminding people to apply those kind of standard journalistic ways of thinking to every kind of information that we come across.
Bailer: So, when you think about the way numbers are reported, I mean it’s often this focus on just a change of an estimate and I’ve you know in one of your- the pieces that we saw as part of preparing for this when you were talking about migration change as an example that you had these point estimates this flux- this fluctuation of 20,000 and whether or not that’s even has meaning, so there’s this aspect of nuance and you talked about both the uncertainty that’s associated with it but it always seems to be to me that there’s huge challenge with communicating that nuance or that uncertainty or look at how much this rate has changed or this percentage has changed of unemployment or whatever endpoint but that may not mean anything. So how do you balance that or how do you tell those stories with that nuance and uncertainty reflected?
Cuffe: Well, what would you have people do?
Bailer: Well, I guess, oh man- turning this around. Robert- you know we’re dealing with a professional here Rosemary?
Pennington: How does it feel John?
Bailer: It feels good. I guess what I would have people do is to think about whether or not it’s a story worth telling. I mean so to me the issue is making sure that there’s an appropriate- I wouldn’t say filter, but caution in place to say that don’t be overly enthusiastic about a point change if it’s well within the scope of what you might expect given the uncertainty that you see in the system.
Cuffe: Yea so the guide- so absolutely. So, it’s not about well sorry, I didn’t mean it to be a test.
[Laughter]
Bailer: No no no no I’m having fun Robert, this is good.
Pennington: He deserves it.
Cuffe: But it’s I don’t think it’s about putting a confidence interval in every headline.
Bailer: No.
Cuffe: It’s about making a judgment that takes into account all of the uncertainty in the data and then writes the most interesting most grabby story that is true to the data and true to the uncertainty. So, the guidance we- I would give on- to our business desk on some of the surveys- the economic surveys that come into the margin of error of it’s not wait until everything is 95% certain. You know you can’t wait until that point but there is a certain range where there’s nothing going on here. Forget about it. There’s a certain range where you want to look at the other indicators that tell you whether or not there is a story here, whether or not this individual change is supported by the preponderance the rest of the data the rest of the information that you’re getting from other sources or people who know the area. And then there’s a range where it’s just screaming obvious. So it’s and it’s adjust your language to match that level of certainty and that’s the gift that the journalists that I work with have is that they kind of once you can get it into that level they’re able to just to write things in just very subtly different ways that communicate how strongly we believe the story and there’s loads of ways of communicating how big a number is you know is it in the headline? Or is it somewhere down the body of the story? Is it something that leads the page or leads the bulleting? Then it’s a more important number than something that’s mentioned later on, and does it have a chart? Does it have a graph? Because charts are realer, you know? So, there are lots of ways of communicating how certain a number is that is not just put a ribbon around it or a bow there’s lots of ways you can communicate that uncertainty and the fun part of the job is getting to that point where you’re working with the people who are great on the words and working together to describe that uncertainty.
Pennington: You’re listening to Stats and Stories and today we’re talking with the head of statistics at BBC news Robert Cuffe. Robert, so early on you talked about how communicating stats clearly for a broad audience was something that you’ve been interested in for a long time, why was that an interest that you developed? because it’s really important that people understand what’s going on around them and risk is a key part of that I mean it’s just crucial I mean and it’s great publishing a paper I mean it’s fantastic if you you know do if you provide some information that you hope will be of use to the general public and it’s not much use if nobody ever reads it or understands it. So, there is that yeah that’s it.
Pennington: Yeah it’s just- it’s interesting because it seems like there is such a reluctance sometimes with researchers and I find that with myself too when I get interviewed about my own work which is not as statistically complicated but that reticence to want to talk about the work that you’ve done and so it’s interesting to sort of always find out what propelled people into wanting to take this on?
Bailer: Yeah this is great. You know what when you’ve looked at the work that you’ve done, what’s been one of the hardest stories that you’ve been involved with trying to tell that really was very challenging for you to do that translation into collaborate with your colleagues in journalism to produce a story?
Cuffe: I’m sorry to be a bit like it might be recency biased but I just think COVID is so hard.
Bailer: I think it probably is more than recency bias, it’s huge.
Cuffe: There are so many different sources of data and it’s so important to people because it’s just turned the world upside down and I don’t know about you but I’m certainly you know kind of looking for an answer. What comes next? And I’m trying everyday to understand that and there was that period at the start where you were kind of almost reflexively checking the phone every couple of hours to look at the numbers you know and I’ve probably gone through that kind of trying to avoid it now for a while you know everyone’s been through that kind of emotional journey with the numbers and it’s becoming more and more politically contested now as we move out of that first obvious hard lockdown and into these more difficult questions about what do we do now as it appears that there’s kind of that we’re, well we’re moving into winter now with a virus that kills people, what are the steps we take that balance the two harms of the damage directly done by the virus and the damage done by the things that we do to stop the virus? Enormously contentious. Difficult from a stance point of view from a volume of data from a communication point of view trying to tie all the different sources together to make sense for people. And also difficult journalistically because we take it very seriously that we want to give who we take impartiality very seriously at the BBC and want to make sure that we’re doing the data right and doing the serious voices in this discussion right even when there’s a wide spectrum of views that need to be examined.
Pennington: There is, I don’t know what the culture is like in the UK, but there has been in the United States certainly a growing skepticism about scientific data and about numbers and a growing distrust of numbers and I wonder if that is something that you have also noticed? I mean, you’ve been at the BBC for three years now, is that something that you’ve also noticed? And how does that if so how does that affect how you approach the work that you do?
Cuffe: I think there is- there are two ways to go around the frequently polarized discussion of facts and numbers. One is to kind of double down on pointing at when things are wrong and the other is to double down on doing stuff right. To you know, put a lot of effort into not just communicating the numbers really well but thinking very creatively about how do you do it in ways that you didn’t have to before? Because some of the debate that’s very polarized it goes on social media and it’s not just Facebook or Twitter which is for old people but it’s on WhatsApp groups and it’s on Tok-tok and so putting a lot of effort into communicating in a way that works on those media. Now, I’m lucky that I work in, frankly, a huge media organization that’s able to do both of those. So, reality check is one of the very big parts of the BBC brand and have played a big role in part of the election in 2019. You go spending a lot of time fact checking debates, fact checking all the press releases that were coming out from the parties during the election and spends a lot of time you know, checking those things, and trying to call rubbish out. But I think we also have a pretty- I think we do a pretty good job of thinking creatively about how do you get the good stuff to as many people as possible? And how do you think about new ways to tell that story well?
Bailer: Can you give us an example or two of some of the new ways that you’re pretty excited about in terms of communicating this type of information?
Cuffe: To be honest I think you’d- well, one of the ways that we have enormous possibility to make data and to make facts grabby now is with the interactivity that our phones give us. So, like a post code look up is just so powerful because it’s not a story about the country or 350 million people, it’s a story about me where I live. And so our ability to personalize stories because we have so much more data available to us now is enormously powerful and the data journalists in the BBC and I work very closely with them I’ve done a lot of work on fun things like which Olympic athlete is the closest body match to you? And down to you know whether house prices have risen after adjusting for inflation for where you live or you know all the stuff on COVID where people are at the moment and a personal calculator on the risk of crime where you are. And so you know ways of making stories personal to people is an enormously powerful way of helping you tell stories and helping you get facts into people’s lives and the phones are a good way of doing it and of course there is a real skill to doing it and because for example and these are only things that I’ve been picking up from the people who understand it far better than I do and well you know you do a portrait, you don’t do a landscape anymore because when as you know when you’re on the bus on the way in to work, you’re not holding your phone sideways, you’re holding it straight up, exactly; you’re scrolling it with your thumb. You need to put subtitles on it if it’s video because you’re probably listening to something else a podcast or like this podcast, obviously or you know-
Bailer: Only if you have really good taste.
Cuffe: Only really good taste-
Bailer: Or actually any taste.
Cuffe: Or some music, so you know you don’t rely on the audio track in your video to tell the story and if you’re doing the kind of interactive lookup that I’m talking about- do you want people filling in 20 things? Because they won’t.
Bailer: That’s a really cool idea that- this idea of the personalized relevance of pieces of news. I mean I think that’s exciting when I think of it in the context of location or geographic index or other components. It’s a little bit horrifying if I think in terms of the echo chamber of hearing certain news stories that might be customized to a certain you know a particular belief, you know I just in terms of filtering news for a certain perspective. I think that that’s not what- that’s kind of the pathological extreme when I think about this kind of customization or personalization of news. Many of us end up customizing our news source now.
Cuffe: It’s something that people are going to continue doing. User organizations are working very hard on how to strike the right balance between putting the important stories in front of people and putting the stories that people will choose for themselves and it’s a complicated technically and editorially.
Pennington: what advice would you have for a journalist who wants to you know be able to report on their community well and use data to do it? You know I’m in a journalism program we send our students to the stats department to take a stats class, but I’m sure there are things that we should be doing to better prepare our students. What should a journalist be looking at or thinking about when they want to be using data in their stories?
Cuffe: thing one numbers are free stories. They’re free stories that your peers and competitors aren’t finding because they are more scared of numbers than you are, and you need- you don’t need very many skills to be able to get great stuff. If you can open a spreadsheet and you can sort it and maybe add up, you will be able to find stories that other people aren’t getting. And I am amazed by you know the journalists who really do that and really kind of hone that skill and it’s not super technical. It’s not necessarily coding, of course, that’s a great, fascinating world in there, but with the publicly available data that’s out there now you can really find interesting stories reasonably easy once you get up to a pretty basic level on Excel, you don’t need to be a spreadsheet guru, you just need to be fairly spreadsheet confident and have a good nose for a story and then you’re away.
Pennington: I like that, fairly spreadsheet confident. I like that.
Bailer: Well, I have to ask the complementary question to this, so help me prepare my stat students to work in this world, to help with collaborating with journalists. You know it’s been one of the great delights of my professional career to interact with my colleagues like Rosemary, but I’m wondering how d I help my students start to get ready to do this kind of work? To be able to think about the challenges that you said such as working with these very tight time constraints and being able to respond quickly to needs and write well and write concisely, what are some things that I can do to help my students get ready for that?
Cuffe: So, well the first thing is just to write. Write lots. You know, that’s the best practice. I think to- so there is a different series of things. If you want to build up your expertise on your own or if you want to take your steps into the world of working directly with the media. Building up your expertise, get yourself in the mindset of news is you know, something that you would schlep up the stairs to your mom, you know? That’s a news story. And there are things that I want to talk about but that doesn’t mean that they’re news. And if my research is interesting it might be interesting to each audience but how grabby is it really and you can do that self-filtering. That’s also a fairly useful way of thinking about when you do have something that is that interesting when you’re thinking about all the subtleties and the caveats and the nuances and the complexity in the data, because it boils down to what are the caveats? Did it change what the headline is? That thing that I would shove up the stairs, you know, can I even run the story? Are the caveats so important that I can’t run the story and how prominently do I need to mention them? You know, do I- you know are they the kind of things that I need to put like in the first three things that I talk about here or can they go down there at the bottom, and if they can go down at the bottom you might need to talk about them or emphasize them more strongly so being very clear about the most important things that you’ve got to say. Statisticians I think in particular may be professionally trained to write because of this therefore I can say that and having demonstrated all of these things I can conclude- and that is the inverse of the way I think that you would want to write if you’re aiming for a lay audience. You need to aim for like why do I care for it? So, focusing on those things and then you can do the baby steps of working with the media. I would say do it with someone like the Science Media Center because they- you know do it with somebody who already knows lots of journalists. Is that fair Rosemary? You don’t want to go out on your own.
Pennington: Yeah, no I think-
Cuffe: You want to work with someone who knows the good ones and the bad ones, and you can translate your first attempts.
Pennington: Yes because I think you know, again, as a former journalist now on the other side I am very careful of who I talk to and I definitely ask around before I talk to anybody so yeah. Going with people you can trust I think is super helpful. That’s all the time we have for this episode of Stats and Stories, Robert, thank you so much for being here.
Cuffe: Pleasure.
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 where you can find podcasts. If you’d like to share your thoughts on the program send your emails 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.