Episode 20: A Statistician Clears the Air
Release Date: 7/22/2016
Barry Nussbaum (@StatisticsBarry ) is former Chief Statistician at the US EPA, and is president-elect of the American Statistical Association. Included among his many accomplishments at the EPA was providing leadership to the agency effort to remove lead from gasoline
Bob Long: The 2016 presidential campaign has been just a little raucous to say the least. We have at least one presidential candidate has promised constituents if he’s elected he’ll shut down the US Environmental Protection Agency. In stark contrast, there are others who have focused attention on the potential harmful impact to children of lead in the drinking water in Flint, Michigan…and possible missteps along the way. For whatever reason, environmental issues such as auto emissions, air and water pollution always generate controversy. And our guest today on Stats and Stories knows a thing or two about getting the ear of a president and having them actually listen to his advice. I’m Bob Long, we welcome to Stats and Stories, a program where we look at the statistics behind the stories, and the stories behind the statistics. And our discussion today is going to focus on the importance of studies about our environment. Before we talk to our special guest, our Stats and Stories reporter James Steinbauer spoke with a Miami University scientist about another water-related hazard besides lead.
James Steinbauer: The city of Flint, Michigan finalized its decision to switch from purchasing clean, treated water from Lake Huron to the Flint River in late April 2014. What city officials didn’t know was that the cheaper, untreated water would also bring dangerous levels of lead to the homes of Flint residents.
The issues in Flint have sparked a national conversation on the safety of drinking water. Much of that conversation has revolved around lead. But Miami University scientist Jim Oris, has spent much of his career studying mercury, and says the environmental and health effects of the metal are just as insidious as lead’s. Like lead, mercury has devastating effects on the nervous system of humans and animals who ingest it. However, Oris says that unlike lead, where half of the metal is excreted through urine and sweat, mercury doesn’t leave.
Jim Oris: It really never goes away. So over your lifetime, you tend to accumulate everything that you’ve ingested. Even though it may not be toxic to an individual fish, over time, because you accumulate it and you don’t excrete it, it can come to toxic levels.
Steinbauer: Oris says this process is called bio-magnification or bioaccumulation. When someone eats a fish that has small amounts of methyl mercury, that metal is never excreted from their body. Every bit that they eat stays with them. So if a person eats a fish with one unit of lead or mercury in it today, they have one unit in their body. If they eat another fish tomorrow with one unit, they now have two units. And so on.
One of Oris’ studies took place on Isle Royal, a national park in the middle of Lake Superior, where he tried to figure out why some lakes tend to have
higher levels of mercury in them and their fish, than others. Oris and his team looked at mercury levels in fish and in sediment profiles. They even took
core samples from the past several hundred years of sediments.
So we looked at historical samples and we looked at contemporary samples to see what trends there were in mercury accumulation in fish.
Steinbauer: They found that up through the late ‘60s, mercury increased over time. But in 1969, around the time the Clean Air Act was passed, the levels of mercury began to decrease.
Oris: So what we found was that you can detect when coal burning started in large quantities through the combination of measuring how much sulfur there was in the sediments and correlating that with the amount of mercury that was in the sediments.
Steinbauer: Oris says that, while the Clean Air Act has decreased the amount of mercury in U.S. lakes and streams, the metal is still a health risk.
Oris: Right now, for mercury consumption in fish, all 50 states have a fish consumption advisory warning for specific kinds of fish.
Steinbauer: Oris says the fish people should worry about are at the top of the food chain — fish that eat other fish, and fish that are old. In Ohio, for example, large northern pike or old largemouth bass would be fish of concern. But tuna, swordfish, and marlin — they type of fish one would find at a local sushi joint — are all fish that can contain large levels of mercury. For Stats and Stories, I’m James Steinbauer.
Long: Joining me on Stats and Stories are our regular panelists Miami University Statistics Department chair John Bailer and Media, Journalism and Film Richard Campbell. And our special guest is Barry Nussbaun; he worked for the US EPA for four decades and he recently retired after nearly a decade as the chief statistician of that agency. Barry also was the branch chief in charge of the successful effort to eliminate harmful lead from gasoline, and he’s currently president-elect of the American Statistical Association. Barry, we welcome you to our show.
Barry Nussbaum: Thank you.
Long: You know, I think adults who’ve been around since back when you started at EPA in the mid ‘70’s, they can remember the belching smoke stacks, they can remember the polluted streams and the littering problems, all those kinds of things that we’ve had…and of course we live in a much…not a perfect world obviously, but a cleaner world than back then. Why is it though…I guess I’ve often wondered why our society just has…or a part of our society has such a negative view of anytime we talk about changing the environment or improving the environment?
Nussbaum: Frequently, Bob, I think the reason you get that view is not that they’re thinking about changing or improving the environment but what might be the downsides. So where they could see an expensive improvement to the environment, they also say ‘we may lose some jobs.’ I don’t think, personally, I don’t think that’s correct, I think we’ve made some great strides and it’s kind of fun to be talking here in Ohio because it was your own Lake Cuyahoga that was on fire that caused some of the concerns, so I think we put that one out Bob.
Long: Going back to that era though, back in the mid ‘70’s when you first got started, I know that you had the ear of President Carter on the whole issue of the harmful impact of lead in gasoline, and I’m just kind of curious…that was back in the early days of the EPA, how you got his attention on that particular issue and how you showed him the studies that proved your point that you wanted to make?
Nussbaum: It relates actually to your first question, you may remember back then we had the Arab oil embargo, we had gasoline lines, we had people having a very difficult time obtaining gasoline. It turns out that when oil was embargoed, crude oil was embargoed, you could actually get more gasoline out of a barrel of crude oil if you allowed a little more lead into it, you wouldn’t have to refine it as much. So here was a question facing the country…should we allow a little more lead in our gasoline, and stretch those supplies a little? Or, because of the harmful effects of lead, should we continue with our lead phased down program? And I was the lucky guy who actually made a graph that ended up on the desk of the president of the United States, President Carter, that showed a correlation between blood lead in children, the actually amount of lead in their blood and the lead used in gasoline. The correlation was incredible, one glance at that graph was like ‘oh my gosh, there’s a connection here.’ And, I know Professor Bailer will sit here and say ‘wait a minute, correlation does not imply causation, didn’t you ever learn statistics?’ I’m well aware of that, that was the same comment President Carter had, he said ‘get me Nussbaum on the phone, correlation doesn’t imply causation.’ The President looked at that, and we actually divided it…blood lead in minority groups, Hispanic and blacks, and they had a more deleterious effect, and he just did not want to be responsible for this harmful health situation, so he allowed the EPA continue to remove lead from gasoline. So it was a point where our analytic work actually made it to the chief decision maker in the United States, very gratifying for a federal employee, but even more gratifying than that, I think health won out.
Long: John Bailer, we’ll go to you for the next question for Barry.
John Bailer: It sounds like there was also the weighing of the type of error you wanted to make in a decision that was part of that. I mean you could keep removing the lead when you didn’t need to or you could let it remain in the fuel when you should be taking it out. So it sounds like President Carter was doing some weighing of decision errors.
Nussbaum: Well, well, yes. He had to balance this public desire to get gasoline and avoid fuel lines versus a health situation, but that’s what’s so critical in our profession. I mean, I’ve spent forty years giving things to decision makers who are always balancing things, and if you can make an effective argument, maybe even more than effective, if you can make it concise and something that they can understand, so communication is the key, that was probably my success in the whole career.
Long: Richard Campbell
Richard Campbell: To follow up on that Barry, it sounds like one of the things that you’ve done is try to explain data analytics to decision makers who have to make hard decisions that often involve weighing a number of things. This is not totally dissimilar to what journalists have to do…they have to explain to the general public in common sense language issues that are often full of data and numbers that are hard to understand. You’ve probably developed some techniques for doing this over time, for talking to decision makers, could you talk a little bit about that and maybe what one of your hardest hurdles was in this?
Nussbaum: Well, you’re absolutely right, Richard, the trick is, and I have a paper on that, it’s not what you say, it’s not what they hear, it’s what they say they heard. What I’ve learned is, and I think that many statistics departments, and the way I was trained… is we are trained to give presentations and to make them fairly clear, but that is only part one. You, as the journalist, are going to hear something, so I hope you’re hearing the right thing, and that is only part two. Part three is when you start to write something and say what I just said. And I remember early on, one of my division directors…I was quoted in the Detroit Free Press and like an excited new employee, there was my name in print, and what he told me was ‘Barry, whatever they said you said, you said.’ I remember that, so my job, Richard, is to make sure that what you write is the correct thing. And what do I use? I use kind of an echo chamber if you will. I will look at the reporter and say ‘so, what did you get out of it?’ and their job, and I understand, they have to write it as concisely as possible, they have to make a point, I know what they have to do, but I will try to get a little echo, try to find out what they think they’re saying, because that’s so terribly important, to me at least.
Campbell: That’s interesting because when we talk about interviewing news sources and when they have a really important…they know they have a quote they want to use, one of our techniques is to say always get your subject to repeat that quote. So, you know that you got it right, and also to slow down to be sure that you get it right, and usually compliment them. ‘That’s a really good quote, I’d like to use that in a story could you repeat that?’
Nussbaum: Oh wow, Richard, so all those compliments…gee thanks.
Long: As a reporter, I remember a lot of times when I talked to somebody who was in the environmental protection business, we often times would ‘I want to make sure I understand this correctly’ so sometimes I would try to make sure that I had it right so I wouldn’t, as you’re alluding to, get it wrong so that you were misquoted in the paper as to what you’re saying.
Nussbaum: And a lot of our stuff is complex. I mean in the environmental…the health relationship, which is why, if I’ve also done anything that is successful I had to immerse myself into what is that all about, what are the health…I mean we were lead in gasoline or automobile emissions, what was the engineering, what was going on here? Let me understand that. It is very easy for a statistician to make huge mistakes if they don’t understand the physical problem going on, so I learned a heck of a lot of nomenclature about automobiles and reciprocating engines and stuff like that.
Long: You’re listening to Stats and Stories where we talk about the statistics behind the stories and the stories behind the statistics, and we’re focused on environmental issues today with our show with Barry Nussbaum, our special guest. He’s currently president-elect of the American Statistical Association and recently retired after forty years with the US Environmental Protection Agency. Along with me on the panel today; Miami University Statistics Department chair John Bailer and Media, Journalism and Film chair Richard Campbell. You know you mentioned Auto emissions; we went through a controversial time back in the…probably mid ‘90’s I think it was where our area was placed in e-check for auto emissions, we had to take our cars in to get them checked. One of the things I discovered that I didn’t know from talking to one of the employees was that there were a lot of these people who were taking out the auto emissions systems, or disabling them, and was that part of the problem back then? I don’t think people realized the harmful impact of the action that was being taken there.
Nussbaum: I’m glad you asked that Bob. One of my very, very first tasks at EPA was to develop a motor vehicle tampering survey. How many people were ripping off emission control equipment, putting BBs in the EGR valve, and maybe for one of Richard’s questions, I had to learn what an EGR valve was, and they said, ‘Barry we think this is going on, how much of it is going on?’ So I worked some engineers, and those days they were you know, add on devices to curb emissions, so the thought was well if you can add them off, or subtract them out, maybe we get the muscle cars back again. So how much of this was going on? So I designed a survey, which was not so easy, we went out to places that did not have an emissions check but had a safety check, and while the cars were there, we had some of our engineers look at it and I devised the sampling, I also learned a terribly important lesson: the statistician has to go out in the field and watch how it’s being implemented. So when I went out there I found out…a big husky guy next to me looking under a hood that he didn’t check the EGR valve for BB and I said ‘why not?’ so he looked at me and said ‘hey buddy, don’t you know that on Fords they put them too close to the manifold, you can burn yourself, you go try it if you like it.’ And I go whoa, there’s a problem here. So you have to see what’s going on. We learned though, in those early days, percentage of cars with at least one emission control tampered with or blocked up with something was nineteen percent, it was very large, that led to a whole new program we had enforcing where muffler shops and repair facilities would rip stuff off, and I think we curbed a lot of that. Although, now the car’s a lot more integrated, it would probably not be so easy to do that today…
Long: I was going to say in today’s world it would be harder to do that than it was back in those days.
Nussbaum: So I don’t want you, after this broadcast, going out to your car and trying to do something.
Long: No, I don’t even think changing the oil is possible anymore; it’s too hard to do. John Bailer.
Bailer: You know one of the things that I’m hearing you say, Barry, is the value of kind of embedding to understand a problem, you know to make sure that you do that. It seems like there’s also an analogue for journalists to kind of embed to understand the reporting of quantitative information. That’s sort of in the same spirit of strategies for succinctly presenting results, I mean, what kind of suggestions would you have for a journalist that wanted to cover a very complicated, quantitative story, or maybe a story that has a lot of modeling of some responses?
Nussbaum: The simple suggestion is to say, sit down and talk to us. One journalist the late Vic Cohn did that and wrote a book on it, I had it autographed, I was very happy about that, came down to an EPA conference on statistics, and, fantastic, I mean, I think we made him the main speaker, but then he sat down and he chatted with us about a lot of health effects and so forth, and he was really striving to learn what’s it all about. Now, I’m well aware that most journalists today, have not just the old, write story and get it in tomorrow morning’s newspaper, but the online, and it’s very rapid, so I understand the pressures they have, but if they can’t spend the time with us then I think we have to spend the time with them, and that’s why succinctly explaining it and saying ‘now, what did you think?’ is probably the best compromise we can do.
Long: Richard Campbell.
Campbell: To follow up on that, are there things that you see journalists do…kind of reoccurring mistakes that just drive you nuts? We talked earlier about the difference between correlation and causation, that’s one that’s come up before, but can you think of an example, or two?
Nussbaum: Oh yeah, my favorite of course is ‘Let’s get to the average.’ I mean, we can come up with confidence intervals, we come up with ranges, we know when presenting things to our decision makers you have to say the what if and how sensitive it is, and if you make a little mistake, does it matter? And then you read the journalist and you know, the answer was 18.46, which is a foolish degree of precision, but it was the average, and yeah, if there’s one thing that irks me, that might be it. The story I always tell…so how do I combat that? The story I always tell about averages is to say that I’ve looked at the data very carefully in the Miami area, not Miami in Ohio, but Miami in Florida where one of my children live, if you look at the data very carefully, you’ll find that the average citizen in Miami is born Hispanic but dies Jewish. And getting people to understand whoa look at how we used average all wrong, sometimes helps a lot.
Long: John Bailer we’ll go back to you for the next question.
Bailer: So, you served as the chief statistician of the US EPA, what does that mean? What does it mean to be the chief statistician of a federal agency?
Nussbaum: Good question John, I always tell people, the Chief Statistician is not the best statistician but is probably the one who can communicate the best. I also tell people when I was named Chief Statistician the first call that came in was from a guy who said ‘Barry, I met you once at a meeting and you may not remember me but, my son has this homework problem and if you have three balls’ you know that type of thing, if you can explain things, that’s what the Chief Statistician does. Chief Statistician represents EPA in a lot of Federal forums with other agencies. Chief Statistician tries to get the rest of our statisticians to be a little more open, they’re kind of a shy bunch, and if I can get them out to say ‘hey we do some good statistics’ that helps. So I think it’s the great communicator. I used to say, cynically, the Chief Statistician is when I flip a coin, they have to listen, but wasn’t quite that.
Long: I do remember an experience I had in a cleanup situation I had in Hamilton, Ohio called Chem-Dyne where they were allowing all kinds of toxic materials leaching out of barrels into the soil and right next to the Great Miami River, that’s where they were located. I remember this guy from EPA was trying to explain what the problem was, but he was doing it in all scientific terms, and I remember just saying to him ‘can you explain, for someone who lives across the street from this what the risk…’ and he did this beautiful explanation and I really learned, and I think sometimes when you’re a statistician you’re used to studying things, you’re not used to putting it in common terms, and I think that’s what you’re saying, that to me that’s one of the tricks you have to be able to do in your line of work, what’s this mean to me as a person?
Nussbaum: Well, there are two things. It’s not just that the statistician has the problem, but being in government work we frequently wordsmith and carefully work regulation language and so forth, so we’re taught to be ever so careful, which is why to you it sounds like ‘wow this is a complex explanation of something.’ So, I have, and maybe as Chief Statistician you get a little more license to do that, I’ve gone out on a limb a little bit and tried to make it more in English and have done it effectively, but you have to be careful because there will be a lawyer who says ‘well you know the regulation really implied such and such and you kind of waltzed over that…’ but sometimes I think you have to do that to help the public understand better. And I think it’s worthwhile that we do, but, as an example, when I worked with gasoline and additives, EPA has about six different definitions for refinery, all for good reasons for what you’re regulating and they make sense when you’re doing the regulation but if I were to look at you Bob, and say refinery, I know what you’re thinking… and I have to realize that when I’m talking to you.
Long: You’re listening to Stats and Stories and our discussion today is focusing on the statistical studies about our environment. Our special guest, we’re very pleased to have with us today, is Barry Nussbaum who worked for US EPA for forty years and recently retired after nearly a decade as the Chief Statistician of that agency. And Barry also was the research chief in charge of the successful effort to eliminate harmful lead from gasoline as we mentioned earlier, currently president-elect of the American Statistical Association. I’m Bob Long along with me are our regular panelists, Miami University Media, Journalism and Film Chair Richard Campbell and Statistics Department Chair John Bailer. I’m going to go to Richard for our next question.
Campbell: Bob just mentioned that you worked for the EPA for a long time, and I wanted to ask you about…that means you’ve worked under different administrations, and in your daily work, how did that impact being… having a Republican administration versus a Democratic administration and also this sort of recent trend…sort of data denier, science denier that we’re in, maybe you can talk a little bit about that as well.
Nussbaum: Sure Richard. The first question about Republicans and Democrats, it turns out, if you had a good analytical argument and a health related reason for doing something, they both listened. I mean if you had to make a global comment, maybe the Democrats wanted to do more and the Republicans focused on just a few things, but they did listen. And if we came up with a good argument…I’m kind of happy to say, I can’t generalize very easily Republican and Democrat. Your second question worries me a little bit more, I think we used to have a lot more of the public behind us and with what we’re doing, maybe that was because there was a river on fire and belching smoke, and we’ve been successful. As a measure of our success, about half a year ago I was in China, I mean some of their air pollution problems, gosh Richard, you don’t need any metering device to know that, you can walk off the airplane and understand it. So, I’m very proud of our accomplishments, but I can also see that as you’ve accomplished a lot, it’s not quite as obvious what’s still out there, and I think we’ve lost a little bit of that public behind us.
Campbell: Do think that’s this sort of knee-jerk, just anti-government feeling that the EPA is part of the government and people aren’t making the fine distinctions about what it’s actually accomplished for us and improved the quality of our lives over time?
Nussbaum: Well, that’s part of a bigger question I frequently get since I live in the Washington area where people want to know what is the government doing, and sometimes it’s hard for me to answer that too. Yes, I know there’s less confidence in the government, certainly a piece of that goes to EPA, and a piece goes to many other agencies, so yeah, we’re in with the pack there.
Long: John Bailer we’ll go to you for the next question.
Bailer: So, if you reflect back on your career, what do you see as the biggest success story of the work that you’ve done and in particular where some insight that you’ve had from an analysis, some statistical modeling that has really shaped a direction for an EPA policy?
Nussbaum: Alright, the biggest impact definitely getting lead out of gasoline. That has been cited as one of the achievements we had that was not just like being on a treadmill, running in place, just to make sure that things don’t get worse. This has positive results; you can look at blood lead levels, much lower. Now, so that in terms of achievement and a personal achievement of course, getting something on the desk of the President of the United States, if you’re a federal employee, that’s as good as it gets. Some of the more fun things: testifying in court. The United States of America vs. Chrysler Corporation and I thought the United States of America was like our Attorney General, it wasn’t. It was some kid out of law school in our division, and it turned into a David versus Goliath thing, which we won. And the reason I loved it as a personal achievement, we had to put together the engineering, the legal, the technical, the statistical, the analytic, the whole package, and very different from other government work, presented to a judge who’s going to declare a winner. I mean this is not writing a memo that some guy’s going to change and might end up in a regulation three years from now, somebody’s going to win this case, and I’m proud to say we did.
Long: Which case was that against Chrysler?
Nussbaum: This was some 1975 model year, pretty big muscle cars that, on the road, this is what made it interesting, on the assembly line they were fine, but as you drove them on the road the carbon monoxide went sky high. But the trick was how do you test cars on the road when you and I are driving them? If I were to go to you Bob, and say ‘Can I borrow your car for a test?’ I don’t know how anxious you’d be…so I had to dream up an incentive to get you to bring your car in. So there was a lot involved and it was a very good feeling to win a case. Now, John would appreciate this, he can guide some of his students, after I testified in this case, about three months, four months later was my own dissertation defense, and that became a piece of cake compared to testifying in court, so there are ways to make that easier.
Long: You know, you talked about your efforts with lead and gasoline, what I think is still shocking to people is the content of lead in water supplies. Flint’s one issue. I was reading an article it was in USA Today I believe, just a few weeks ago about the number of schools and things like that that had high lead content in water. You mentioned yes, that the world is cleaner but we still have a lot of issues like that. How much of a problem do you think it is the lead content in water, because, again, children are often cited frequently?
Long: Unfortunately Bob, it’s still a big problem. I naively thought when we got lead out of gasoline, and that by the way came after we got lead out of house paint, and these were the two big problems. The trouble is there were a lot of older houses where kids would still eat lead, I am told, by the way, that lead chips are sweet; there is a reason that children eat them. I didn’t know that, I don’t try it on my own. So I thought the lead in paint and the lead in gasoline would solve the problem, it turns out between older housing and this terrible situation in Flint, Michigan, I mean lead is there, it is a big problem, and apparently it still is.
Bailer: What advice might you have for training the next generation of statisticians or journalists that want to work with decision makers and government and beyond, but also environmental science in general?
Nussbaum: For the next generation, and I wouldn’t say statisticians and just journalists, I’m concerned about lots of people who do analytic work who have access to big statistical packages, R or something like that, who don’t call themselves statisticians, or who don’t call themselves journalists, but do a lot of analytical work, I would like them to do it correctly. I would like them to explain it succulently. Then I think we can put our combined resources, and remember my emphasis has always been on what is the problem that you want to solve? So immerse yourself in that. The joke I used to tell was…I mean lots of people would come to me as the Chief Statistician, and either they would say ‘My boss told me to come in, what should the sample size be?’ and I said ‘Do you mind telling me what the problem is?’ ‘Well they told me I had to come here and get the sample size.’ And much like the, this will date myself, but the, I think it was 1967 to ‘68 movie The Graduate the word was ‘Plastics’ I would say ‘Stratify’ and they’d never thought of that in terms of ‘oh, here’s something that’s important.’ So, it’s the communication, I think between what’s you’re doing. The other problem was, many mangers would say, we had an analytic piece of work, did you run it by Barry? And the answer is ‘Yeah, we went right by Barry.’
Long: Richard Campbell, we’ll go to you for our final question for today.
Campbell: This follows up on John’s question about what can we do as educators to help train students better, and this is a personal question, you got into the business of having to explain to the general public and to decision makers, are there things that you wish you would have had as an undergrad that helped you do this kind of work? Or were you just sort of self-taught, after a while you just became a better communicator and you sort of figured it out on your own.
Nussbaum: I was self-taught. I was forced into it, I remember the first or second day, they started throwing environmental terms, my favorite one was SIPRevision I didn’t know what it meant, I went home and I looked, I didn’t know if it started with a ‘C’ or an ‘S’, I looked it up, it turned out it was ‘State Implementation Plan Revision, I had no idea. And it hit me right away, ‘Gosh Barry, you had better learn their nomenclature’ and if you do, you know what? They’re going to lean over and learn some of yours as well, and that cooperative, collaborative model helped me for years and years.
Campbell: Very good.
Long: Barry Nussbaum, we want to thank you very much for joining us and sharing all of your wonderful insights about the US EPA on Stats and Stories today.
Nussbaum: Thank you.
Long: If you’d like to share your thoughts on our program you can send us an email to firstname.lastname@example.org be sure to listen for future editions of Stats and Stories where we always try to focus on the statistics behind the stories and the stories behind the statistics.
Click to close the script.