Dr. Regina Nuzzo is a freelance science writer and professor in Washington, DC. After studying engineering as an undergraduate she earned her PhD in Statistics from Stanford University. Currently she’s teaching statistics in American Sign Language at Gallaudet University, the world’s only liberal arts college for deaf and hard-of-hearing students.
Dr. Nuzzo is also a graduate of Science Communication program at the University of California-Santa Cruz. Her science journalism specialties center around data, probability, statistics, and the research process. Her work has appeared in Nature, Los Angeles Times, New York Times, Reader’s Digest, New Scientist, and Scientific American, among others.
Episode Description
What is a median? How about an interquartile range? Don’t even get me started on how to define a p-value. These statistical concepts are hard to grasp for your average statistics student, but imagining how these types of definitions translate into American Sign Language is a whole other ballgame. That is the focus of this episode of Stats+Stories with special guest Dr. Regina Nuzzo.
+Full Transcript
John Bailer
What is a median? How about an interquartile range, a p value, don't even start talking to you about p values. We don't want to get a fight going on here. These often feel like very abstract concepts that are difficult to imagine and, and also difficult to communicate. There are aspects of painting these concepts with words that are a critical part of this. And that could be whether you're teaching a concept like this and American Sign Language, or you're translating these ideas into a comic, for a larger community, a larger audience. I'm John Bailer. Stats and Stories is a production of Miami University's Department of Statistics and media, journalism and film as well as the American Statistical Association, Rosemary Pennington is away. My guest today on Stats and Stories is freelance science writer and Gallaudet University professor Regina Nuzzo. She has a PhD from Stanford and a graduate degree in science journalism from the University of California, Santa Cruz. And her writings have been published on science and medicine, health statistics and the process of scientific research in a variety of outlets ranging from Scientific American nature, ESPN and the Los Angeles Times as well. She now teaches statistics in American Sign Language and has co authored a comic, I mean, don't go figure. I mean, you know, a statistician involved in a comic that does the same thing so well, Regina, welcome to the show.
Regina Nuzzo Thank you, John, thanks so much for being here.
John Bailer
Oh, it's great to have you and and, you know, I'm sure I didn't do justice to the richness of ways that you think about these ideas. But I did mention a couple of things that you do routinely, in terms of communicating statistics, one has been translating concepts to a comic, as I mentioned at the end, but also in terms of your teaching and American Sign Language. Could you could you just talk a little bit about some of the challenges of making these abstract concepts into kind of visually communicated ideas?
Regina Nuzzo
Yeah, sure. So I think my background teaching at Gallaudet teaching in ASL really shaped how I think about communicating statistical ideas. So ASL is a very concrete language. So American Sign Language, it's up in the air, I always think of it as being four dimensional, because you've got the three dimensions of space and time. And to me, English, written or spoken, English feels very linear. It's word after word after word. And ASL doesn't really work that way. It's very spatial. And it works best when you're showing pictures when you're showing action, actually. And I was not fluent in ASL really, before I started teaching a go at it. So I had to learn how to do this, how to take these statistical ideas and throw them up in the air.
John Bailer
Oh, wow. I love that image. And, you know, it's, I guess, it's ironic that we're talking about the the spatial visual communications on an audio podcast, you know, apologies to the listeners, but you know, I guess the best thing to do would be to put on some Pink Floyd at this point and, and listen to the show as well with that background just to get yourself in this mindset. So, Regina, could you help with an example of, of an idea and and paint a picture with words of how that was communicated?
Regina Nuzzo
Yeah, some examples of that, um, let's say we want to talk about correlation. Okay. So the word correlation? Okay, that's not kind of interesting. You're talking about two things that relate together. But how can you make that visual and the choice that I'm using to make that visual really can shape how people are thinking about correlation? For example, am I showing it as a scatterplot? Right, am I showing a bunch of dots? And kind of using that as the basis for correlation? Am I showing a straight line or a curvy line? So when I arrived at Gallaudet, the typical sign for correlation was something that was kind of involved. It was akin to the word for just relationship, or, or association or togetherness, combining. And they didn't really convey this idea of a one linear thing. So I teamed up with a Deaf Studies ASL expert, and one of my statistics students and so you know, guys, let's get together and talk about what correlation really means. What's the best way to throw that up in the air and show it and we developed the You signed for it. Yeah. And I feel like the words that we use help make these ideas more solid in people's minds. And so I'm proud that that's just one sign. I'd love to do that with all this. That's vocabulary that we've got
John Bailer
Could you describe that sign for us? But you know, that gets in that spirit of painting? The picture is that
Regina Nuzzo Oh, no. Now, as an elder toward a picture worth 1000 words, oh, he is signing it. I'm using both my hands. So it's talking about two variables, okay. And I have them coming together and pointing together. So they're coming from two different areas in space, and then coming together, and kind of making direction. And the nice thing about ASL is that you can easily modify things. So if I want to show that it's a positive correlation or a negative correlation, then I can do that just by changing the position and the direction. Exactly, exactly. So. So this is nice. The old sign was something that didn't really convey the directionality. And it also didn't convey this idea of a straight line. And I really wanted my students to realize we're talking about linear correlation.
John Bailer
Okay, so now you're not imagining, okay, how you've talked about directionality that you've captured in the sign that you've developed? How about this, the magnitude, the strength of this linear relationship?
Regina Nuzzo I wish I could do that. You know, that's very, oh, now you're getting me out.
John Bailer
So now I'm thinking about whether they're right. How much? You know, we're just, we're Regina and I are just making all these gestures to each other and they're all kind and productive, by the way, so.
Regina Nuzzo
And they, right, and they're very good, right? So you can, can you actually show that in ASL through kind of how much you're emphasizing how much kind of muscular tension you're bringing to the sign so you can convey the strength that way? Yeah.
John Bailer
Like, there's very much, there's a motive component to the gesture that sounds really, really fascinating as part of the way in which that's conveyed.
Regina Nuzzo
It didn't you know, that frustrated me when I first started teaching Gallaudet, this idea that ASL had to have emotion with it, because you know, x plus three equals five, there's not really a whole lot of emotion going on. And they are, well, I'm surprised that it's five. The more I understood the language, the more I saw the kind of the parallels with statistics and math. It's a very movement oriented language. So when you think about, you know, x plus three equals five, it's x, but then you're taking something else you're taking, you have this unknown in your hand, and then you're bringing three things to it. And then it becomes five. So mathematical equations, in some ways, are verbs. There are a lot of verbs in there.
John Bailer
Oh, I love that. I love that image. If I just continue this, this idea of you helped me with thinking about correlation and sort of the gesture, the sign for that, and also the directionality there, and possibly the issue of kind of how the, the, the magnitude of it is also conveyed. We often talk in an intro class about, you know, this, this old, the old saw of correlation is not causation. And so when you go from how you've sought you think about signing a story, that's that correlation is how do you kind of connect that to this idea of causation?
Regina Nuzzo
Yeah, you know, this is a great question. And that makes me want to sit down with my collaborators again, and we can come up with a better time for causation. Because the typical time for causation is not really well set up to show the difference between that and correlation causation, it's a little odd, but it would be nice if we could set up something. Well, we have correlation. And then we have causation on the other side, and you're showing the difference. You're showing how causation is one thing leading to another, a force in another and this correlation is two things just happening together at the same time.
John Bailer
So I'm curious now, when you were thinking about developing this new sign for correlation, how did you evaluate the effectiveness of one sign versus another in terms of communicating that idea?
Regina Nuzzo
Yeah, this is where I really relied on my collaborators there because we both have native signers. And just like if we were to make up a word in English, there are certain laws Other combinations that make more sense than others. And so they have this, this deeply embedded sense of okay, that's your persona. And that's not where you were, how you move your hands together. A lot of that hedges internal rules that I don't have. But some things are flexible. For example, we were talking about how to talk about median. And where do you have a number list? If we were to talk about a list of numbers? is the smallest number at the top? And then the biggest number at the bottom? You know, top to bottom? Or is it? Or is it the other way? Or are we talking a number one left to right? Does that make sense? What I'm saying? So where? How did our brain like to picture a list of numbers? If we're talking about median, right? Are we picturing it top to bottom list left to right? A cloud of numbers? What is the image in mind? What image do you have in mind? When I say median?
John Bailer
I almost always have this as I think of my mental model is that of a teeter totter? In a playground? Yeah. So um, so probably the number line where there's, there's a very massive being placed on that, right. And so for me, the image that comes to mind is that a, kind of the, this balance point, fulcrum component to it. But of course, then the median is an interesting point, when you're thinking about kind of within an ordered set, where you play the middle, middle value, so one is balanced, and the other as middle and rank set. So those are, again, that's an idea of the abstraction of these concepts into something that we can represent and communicate, or you can actually show
Regina Nuzzo
Yeah, same thing means, right? I mean, it's gonna be that balancing point of a histogram. So can we create a sign that really conveys that? Yeah.
John Bailer
So what's, what's the hardest concept to communicate in an interest that class with we're in ASL?
Regina Nuzzo
Well, peak value, let's say just p value? It's hard to convey no matter what. These things rely on these counterfactuals? You know, statistics relies on counterfactuals, which is abstract, because suppose you didn't know what actually happened. All the other things that could have happened that's difficult in English, but it's definitely difficult when you are trying to make something visual, because you're inherently talking about things that are not visual.
John Bailer
Yeah, there's this dependence on knowing that something is conditional on something else. And that's there's a lot of it as you kind of deconstruct it, there's a lot of subtlety.
Regina Nuzzo And in this it, well, think about conditional probability, the probability of A given B. So given given what, given right at the very English thing, so if you already know B, now you enter this world already know B. Now with the probability of A.
John Bailer
You're listening to Stats and Stories, and my guest today is statistician and science writer Regina Nuzzo. So you know, we've been talking about communicating abstract concepts, using ASL and you've talked about the development of new signs. You've also been involved in projects with a very different type of modality of presentation. And I really love the comic that you worked on. So I was wondering if you could talk a little bit about, you know, what the comic was about? And then maybe we'll dive in a little bit deeper to figure out what led to it in the first place.
Regina Nuzzo
Thanks. You know, this comic is one of the favorite projects I've ever done. It was for magazine club, knowable magazine, which is the lay language or public facing publication of annual reviews, and annual reviews, publishes reviews, nothing journals, or reviews and various fields. And I had worked with the editors there before. And they contacted me and said, Hey, do you want to write for us? What after trying something different? We'd like to try a comic? Are you up for a comic? And anytime an editor says, Are you up for that, then? My answer is always yes. So I have never had any artistic ability at all. But I really love the idea and the challenge of taking some statistical ideas and making not only a story out of them, but making them visual, very visual. So I looked through the various reviews that they had there for me to pick and find a comic adaptation and I found a paper by Howard Stern. And it was forensic issues, statistical issues in forensic science. And I thought, oh, forensic science. Well, that's fun. You can bring up Little bit of a crime thing in there, you can have a little bit of a story. And I thought I would have at least a chance of talking about applications and visual things in statistics around this, this particular idea of forensic science. The problem was that it was talking about likelihood ratios.
John Bailer
Yeah, at this point, you've lost the audience.
Regina Nuzzo
Likelihood Ratio? How are you going to make that concrete? Difficult concept?
John Bailer
Yeah, I love likelihood. So come on, Regina. They're awesome.
Regina Nuzzo They are awesome. Well, what I love about them when I get down to thinking about it is that they're their ratio. So it means one thing to another, right, you're comparing two things that is the top bigger than the bottom, is when it comes down to. And when I thought about it, I realized, Oh, hey, that's kind of fun. Because you can have this where there's old fashioned legal scales. And a ratio is kind of that it's weighing that top and weighing the bottom, you have weight of evidence on the top and weight of evidence on the bottom. And that what those those old skills would do. So I decided that would be my visual metaphor.
John Bailer
Was that the first metaphor that came to mind? Once you selected this idea of doing forensic evidence?
Regina Nuzzo
Oh, good question. It was not that I actually considered having some sort of black box, or like a machine like a factory thing. And that would have been the easy way out. But also, I didn't like the idea that it kind of perpetuated I think, to non statisticians. That statistic did a black box, you put it in and the magic happened, and then out pops the answer.
John Bailer
So what was being compared in your story? What were the two that competing states of nature that were compared as part of this likelihood ratio, evidence for prosecution and evidence for defense?
Regina Nuzzo Okay, talking about likelihood ratios in forensic evidence, you're saying, Okay, we have a look at this DNA match. Right? So what are the chances that you would have this DNA configuration of this DNA match if the person were really innocent, versus the probability that it would happen if they were guilty?
John Bailer
So what was the crime? I mean, I feel like I'm, you know, turning into my favorite mystery series now.
Regina Nuzzo
You mean, the crime that I
John Bailer
Yeah, what was the crime being investigated? In your comment?
Regina Nuzzo
Oh, my. Well, the first thing that I had to decide was what type of evidence are we looking for? So in his paper talked a lot about fiber fragments, you know, fiber evidence from clothing, or glass. So I tried, I actually had little thumbnail sketches. Can I make an entire comic around a thread? I really have no way to make that visual. And then I tried it with glass and that was not going to happen either. So I made some footprints, which is, okay, shoe print, which is actually easy to see. And then I knew I wanted to make it my villain to be a little non-traditional. So I chose a woman named Scarface Diane. And the shoe Pruitt was a Birkenstock sandal.
John Bailer
As most of the bank robbers you've known over the years have worn Birkenstocks. Is that it?
Regina Nuzzo
That is it. So I actually left the crime, a little unspecified. I didn't want to get all macabre and have a murder. So there's some sort of crime. But the important thing is that the crime, yes, this the thief, the suspect war, Birkenstocks.
John Bailer
I love that you that you were, as part of the the formation of this story, this comic that you had, while you knew you were going to be comparing evidence giving, you know, giving guilt or innocence, the nature of the evidence, what could be visually communicated clearly, and accessible was a major part of what had to be decided. Were there other kinds of decisions that you sort of were similar to? I mean, you did identify the villain, and what other kinds of decisions needed to be made as you were putting together this comic?
Regina Nuzzo Well, just like in any feature story, you have to give a little background. So I had to give a little background on the idea of what's happening in forensic science. Why is this even an issue? And I had to decide how I would present these presidential commissions and reports and national academies, how was I going to present that visually? And I ended up doing some creative things around that. Just where, I think at one point, I had a skywriter. You know, carrying a banner through the air and right, yeah, that's the nice thing about comics is that you can get all kinds of fanciful and whimsical in there.
John Bailer
So just to give us some sense of, of scale and size, you know, you've written, you know, kind of General, General pieces for sort of science writing, and you have a certain number of words that you have, how many words what how does that compare to a comic that you might be that you worked on? And here's the length?
Regina Nuzzo That is a great question, I had no idea of the rhythm of a feature, I know the rhythm of a written feature. So I know about how long I spend on the lead and about where the graph comes in and, and how long you're going. And I didn't, I didn't know that I had to sketch it all out. So I had an idea of how many, how many panels, how many rows would fit in a page? And how many pages? So I tried to just, you know, divide it up mathematically and say it shouldn't be about here is that there, and I had the rhythm all off when I first got into it. But that was just like when I was starting to write it, too. I didn't know how to do that. But first, I want to answer your question, because I feel like I didn't do justice to it before. What were some of the decisions I needed to make to make things? Visual? I think one of the most important decisions I had to make is how am I going to communicate statistics and probability there? What could we talk about oh, statistics to the rescue, or, you know, statistics to this. So I decided to make them superheroes. So I kept wearing statistics with a big ass and I kept wearing probability with a big P. And they came crashing through a wall and they had a briefcase full of, you know, they opened it up, and all this great stuff came out. And you know, our code was in there. And I really wanted to bring in the idea of statistics to life, the field, you know, the concepts in there. So they became superheroes.
John Bailer
I love that. I sometimes talked about in an intro class, that one of the things that students are learning is data, self defense, you know, maybe having a superhero is even better to have a better image to have in mind. I think that's wonderful. So, I'm just curious when you had to. So you were using this piece that had been published annual reviews related to forensic science. You were working on their translation of that into a different modality. The comics, were the constraints of comics in terms of pages and panels, determined the kind of text that you had, but then even within that you had to find rhythm in terms of the presentation. So once you did, did you have any ideas about the storyboarding of each of those panels before you enlisted? The assistance of an artist?
Regina Nuzzo I did? I did. Again, I am not an artist at all. So my story boarding was just trying to think in terms of panels and what would happen, what was the action happening in each and panels not quite like a paragraph, I would say it would be like a chunk of paragraphs, where you're kind of conveying one point, and then you're moving on, and then moving on to the next. So they need to have some sort of separation in a way that is a different sort of a different sort of beat. Anyway, so I sketched it all out with my horrible stick figures. And just to see, was I taking too long? And some things could I even hope to make something visual? How am I getting across the idea of a likelihood ratio that's too big, you know, when is enough evidence there? How am I talking about the weight and the scale? All of that I had to sketch out before I brought in my magical artist.
John Bailer
Well, and you said that you have this experience of what the rhythm of a written feature would feel like? And you know, it seems like when you're writing with that, with the paragraphs that you're writing, you have a certain sense of how much punch per paragraph Do you have? It seems like the panel has to have a lot of punch, you know that because they're just there won't necessarily be as many of them and that punch has to be integrated with the image that you're presenting. So I find that really fascinating to think through that can. Can you give an example of how you kind of worked with both the images? You know, you've already given us the one that scales the scales of evidence for the ratio. Where was there another picture that you can paint for us from the comic that you really loved?
Regina Nuzzo So there's one section that I felt like I needed to work on to really get the rhythm. And this was trying to get across the idea that a likelihood ratio, and like hypothesis testing, traditional hypothesis testing, set up hypotheses and pitch them head to head. And, you know, not just trying to knock down a null hypothesis, but setting up the null hypothesis and the alternative hypothesis, right, the defense and the prosecution, and this idea that you want to try to knock it down. And whichever one you can't knock down, and I ended up using a punching bag, you know, one of those old clown, the inflatable clowns, and the defense attorneys trying to punch it down, and the prosecution attorneys trying to punch it down and bring to life that, that, that idea, and that took a lot of work to get the rhythm of that just right.
John Bailer
So I'm just curious, what has the feedback been like that you've had to the comic? What do you know, what kind of reactions have you heard from it?
Regina Nuzzo
People love that, you know, one person posted on Twitter and said that they were going to make this their Halloween costume, this probability in statistics superhero. And I feel like that's as much as you could hope for in life, you know, as a science, that journalist is to inspire someone's Halloween costume. People saw that it was fun because it got these ideas across, you know, in this different way. And it had a lot of humor in it. And in the appreciative that
John Bailer
I thought it was fascinating and a huge success, and in fact, recommended it to a number of students and colleagues. So I'm one of the fans too, although I don't I don't have my superhero costume though. So I have not yet reached that level of fandom. Sorry, Regina.
Regina Nuzzo
Maybe we should get you a cape, I think Oh, zero. All on your own anyway.
John Bailer
Oh, man, what a silver tongue devil. Well, that's all the time we have for this episode of snaps and stories. Regina, thank you so much for joining us today.
Regina Nuzzo Thank you.
John Bailer 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 can 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 discuss the statistics behind the stories and the stories behind the statistics.