Quantcast

Wednesday, January 8, 2014

15

Interview with Susan Murphy, Genius Grant-Winning Statistician

Susan Murphy is a statistician developing new methodologies to evaluate treatments for chronic and relapsing disorders like depression and substance abuse. In 2013, she received a Genius Grant from the MacArthur Foundation. We talked in her office before the holidays.  

So you grew up in the South?

Yes. About half the time in central Louisiana and half in southern Louisiana.

And you were a math major at LSU. Were there a lot of other women in your classes?

I don't remember any other ones.

None at all??

No.

What was that like?

You know, I'm not sure if I was even that aware of it. This might be common to women in my generation who have done well in the sciences, but I truly don't think I had the social awareness to feel self-conscious about it. We were the pure math majors, and it was some guys and me, and that was it! Certainly now I do notice discrimination. After graduate school I started noticing these things more and more. But, at that time, I don't think I was clued in.

It's kind of nice to hear that you didn't notice any side-eye that people might have been throwing at you. 

Yeah. It is sort of sad, too! I do think the minorities that did well in that sort of a culture, at least back then, tended to be people who didn't pick up on social cues. Maybe people said all sort of things to me! I would have had no idea.

I was that way with racism growing up, and as a result I've always thought there's something to be said about being shielded from discrimination by any means, even if it's a lack of awareness. Less time engaging with discrimination, more time working. 

Yes. And now, really, I do want to be socially aware. Now I certainly want to know if people are comfortable, how they're responding to what I'm saying. But at the timeeven the guys I was with, now that I think about it, were really the same way.

So you always loved math? I assume you took to it really early?

Yes. Well, I don't know if I could say that I loved it. I'm from the South, so I was being trained for a particular kind of life. When I was in high school, I made all my clothes at home, I took advanced home ec, which is very difficult, almost like an art. I was being trained for that sort of thing. It's just that I was also good at math. I have a sister who's younger than me who's very bright, and she ended up with an MD/PHD in physics, and so in our family being good at math was just normal.

Would the life you were headed for have allowed you to have math as a career?

A career, sure, but not an academic career. In college I started off in accounting, because I thought accounting was math. Turns out it's not. So I transferred out after a semester.

What made you decide to go to grad school for statistics?

After college I went to Germany to study group theoryand, while I was there, I also took a statistics course and I thought, oh, I like this. I wanted to do something that would have some near-term applicability. I wanted to be a little more applied, to work on something that would help me help people. At the time I didn't realize it, but I think that's very feminine.

Still, the really cross-disciplinary stuff I do, that didn't come until later. Even in graduate school, I focused on the most mathematical part of statistics; even in the first six or seven years of me getting into academics, I stayed in that mathematical realm.

And not that many people are there! I was doing some reading and saw that only 25,000 people in the US are pure statisticians. I think it's a job that people (or at least I) might not understand very clearlyI think I get statistician-adjacent jobs, like actuaries and financial analystsbut how would you explain the sort of career path of a statistician to someone unfamiliar with the field?

Well, let me think. It's hard when you first get out of grad school, because you really have to figure out what problems are relevant problems to work on. It's easier to work on hard problems that are more abstract, particularly if you are math-focused. So for me, I got a postdoc with a guy in the Netherlands and I went to him and said, "Do you know of any open problems?" He said "Yeah, I've got this one, I haven't been able to solve it."

And I solved it for him. It took me seven months.

That's really what I did. I went and looked for problems that no one had really solved. And that's one of the great advantages of a quantitative field, especially for people who are minorities. It is a great boon to have no one be able to question your successesthe problems are black and white, solved or not. You got it, it's your proof.

So there's a concrete sense of achievement.

Yeah. And so, the guy in the Netherlands wrote me a letter to go to the Mathematical Sciences Research Institute at Berkeley, and I solved another problem, and that got me to the next place. And that's how I played the game.

What sort of problems are you talking about?

They are related to real life but abstracted away from life. Mathematized. That first problem, in the Netherlands, it was related to these studies on twins where they were trying to disentangle genetic versus environmental influences on how long each person would live.

Were you interpreting data?

There was no data. I mean, really abstracted. Alpha, beta, gamma! But if you could solve this kernel of a problem it would have a very wide application. Of course, after awhile, I realized, this was all fine and good, but not how you make a big impact in real life. So I moved more towards trying to mix theory and application.

How long have you been working on the research area that got you the Genius Grant?

Well, while I was doing all that math stuff, I started going to these brainstorming sessions. This was in maybe '97. I went to a researcher who was working with little kids with conduct disorder. She had her own problem: the data seemed to indicate that the more treatment these kids got, the worse they seemed to do. I was really interested in that. Why was this happening? The data was so odd. That's how I got interested in this.

I'm really interested in the SMART (Sequential, Multiple-Assignment, Randomized) trials, which you explained a bit in that video for the MacArthur Foundation. So, as I understand it: for chronic illness, you have to make not a single medical decision but a complicated series of decisions in treatment, and what you're working on is a way to test and improve these decision-making models?

Yes, that's great. So if you think about how clinical trials were developed, they were there to evaluate whether or not a single medication cured you; there was no need to think about what was next needed for the patient. So after one treatment, you often didn't have to worry about the patients anymore. But this way of viewing clinical trialsas evaluating or confirming one treatment versus the other as a single-stage treatmentit makes a lot of sense if the majority of our diseases are terminal or infectious diseases, but not so much sense today, when so many of our diseases are chronic.

So, when a clinician sees someone, he or she has to be prepared to help the person over many cycles. But our clinical trial setup is still not formulated to adequately inform that sequence of decisions. Let alone mental health or substance abuse, even cardiovascular disease is like this—as with the kids with the conduct disorders, there are lots of possible treatments, the treatment will have to be ongoing, and people may respond differently. Over the course of their career, clinicians are making these complex decisions: they've seen patterns, and they use their instincts to make decisions, but you want there to be some data there to help them.

I thought it was interesting that you're dealing with a lot of stigmatized conditionsalcoholism, depression, etc.

That may be more of a personal interest than a large research interest. I don't think cardiovascular disease is stigmatized very much, and that's much more prevalent than schizophrenia. I've just always been interested in brain disorders.

Yes, so you also have a position in the psychiatry department at UMdid your interest in brain disorders develop in tandem with your research?

It probably started with my conduct disorder research. As you say, people who suffer from these things are often maligned, and that appealed to meI would like to help these people. And it is also a major frontier in terms of science.

Additionally, another interesting thing about the clinicians who work in this field is that they really understand the need to change one's behavior, to motivate people, to keep people engaged in their treatment. They understand the importance of context: family structure, where a person lives, their stresses. Traditional fields recognize this but not to the same extent, although it's really the same everywhere. If you're not engaged in your treatment for blood pressure, you'll forget to take your pill.

And then, mental illness is even more complicated...

They're multifactorial. Many things can encourage maladaptive behaviors. You can be exposed to a trigger substance, you can be not feeling well, you may not have a lot of social support, you may have a lot of stress. And all of this is just as important to cardiovascular disease, it's just that with that, there are medications with clearer effects, things to focus on that allow you to sort of let the behavioral factors go. But those behavioral factors, in the end, may be even more important.

The stuff you've worked on, what's the ideal application of it 5-10 years down the line?

So lately I've moved a little bit away from the SMART trial. The SMART trial is about accumulating data to help clinicians help you; now I'm accumulating data to help you help you. This has to do more with mobile devices, phones. Let's say you're trying to recover from substance abuse. This is very, very difficult. You are often having to learn how to deal with new stresses. How can we deliver interventions via your phone to help you manage momentary stress and keep you from going down the block and buying some coke?

I would be very happy to have had a meaningful impact on substance abuse and recovery. What I want, in ten years, fifteen years, maybe, is to have an intervention out therea coach on your phonethat will learn from your responses and create evolving strategies to offer you. Because hard moments in recovery require different strategiesdo you need to be distracted? Do you need to confront your craving? Do you need to check in with a real person? And the strategies in the first six months might be very different from the strategies that work in the second half of the year, or at a year and a half. Your life changes; you want the strategies to change and remain effective.

Do you think this would be better than what a person would deliver?

I don't know if we could ever replace that. But a phone might put you in contact with a person, and you won't have to wake up the phone in the middle of the night.

Basically, this is against one-size-fits-all.

Yeah. And I don't know about you, but I've tried some of these quantified-self apps just out of curiosity, and it's hard to maintain engagement. They get boring.

So how would the app know what you need?

There are a lot of big issues to be dealt with. First of all, the big challenge is not wanting to place a burden on you. I cannot tolerate my phone bothering me, you know? So you want the phone to collect as much information as possible, but passively. It might collect how active you are. I carry my phone around all the time, so it would know what I'm doing, if I'm near the liquor store, if I'm near a place that tends to be a problem.

It'll perhaps nudge you towards one pizza if you're ordering five of them at 3 AM.

It'll know who you're around, even.

Or if you're alone for too long!

Yes. And of course the person has to want this. The person has to opt in.

The important thing is to figure out how to have the phone only doing things that help you understand your disorder better. It should not be burdening you. It should be there to make you more self-aware of your actions and what they meannothing that would raise a stress level, but something that you could employ over a long period of time.

So you teach undergraduate and graduate students, you're applying for grants, you're mentoring students, you're conducting researchdoes all of that work seem cohesive, or are you spread over a lot of different spaces? How do you balance it?

I try really hard to have everything fit into this major project. My undergrad teaching, of course, has nothing to do with this. But my graduate teaching, I try really hard to have it relate to this focus. I guess I can never really succeed at this balance, I can only really strive. If you are able to bring in funding, there's ways that you can then allocate more of your time towards that research.

So how will your research change now that you've gotten the grant?

I think it'll change a good bit. Mainly this will lead to additional interesting collaborations. A lot of people think about it in terms of the money, but it's really not the money. The money goes away really fast. Graduate students are expensive to hirewe need to provide them health care, salaries.

And it’s good that you are providing them with both! How many students do you have working with you on this project?

I have four grad students and one postdoc. And it is wonderful to be able to take care of my students, no two ways about it. But the real thing that the MacArthur Foundation gives is that it opens the door to different collaborations.

Is statistics taking more of a collaborative direction today?

Yeah, I'd say for the last 10 or 15 years the field has really been moving that way. Which isn't to say that we don't like the abstract things, the pure mathematical things. But I think we should always be solving a real life problem. And actually, even when people are working on very abstract areaswhich is still great workthey often look to people in the field at a different end of the spectrum to say, how can we use this?

I saw in that video, you were talking about getting the phone call from the MacArthur Foundation, and you said at the time you just sort of slid down onto the ground.

Yeah. Very much in shock.

Do you feel additional pressure? Has the joy of getting the grant stuck around?

Yes. The joy is great. But in terms of pressure—this semester I've been spending a lot of time on my teaching, so I haven't had a lot of time. I haven't had any days off now, even weekends, for a couple of months now. Four months. I haven't even had time for it to sink in.

Wow. That is a lot. Do you have a lighter semester next year?

Yeah. And I think it'll start to sink in then. But there are just so many things to do every day that I just can't even…

It's a good reminder, though, that you don't get a Genius Grant by just kicking around. 

Yeah, but I also don't want to be working 7 days a week for too much longer. I worry about my research, my ability to think out of the box. Anyway, my schedule is my fault—I was trying to do something really interesting with my grad course, I should say, it's not Michigan making me.

Does being a statistician make you read the news differently? Are you always looking at polling or study reports and thinking that things are poorly designed, or interpreted incorrectly?

Oh yeah. I see alternative explanations for results in the news all the time.

Does that bother you?

Well, I've got this viewpoint—and I don't know if this is very mature—but I think it's a big game out there, and we all have to be prepared to play it. Everyone is trying to frame things in their own way, and we all have to try and be as educated as possible so we can understand the degree to which it's a game, and know if someone's trying to pull the wool over our eyes. Even if it's someone I agree with!

That's why I wish more logical thinking was taught in high school. Many people will never use algebra again, but logic and statistics, you can use that in real life. You need to be critical about new drugs coming out, about numbers you see. People understand that critical thinking is important from a qualitative point of view, but it's just as important from a quantitative point of view. I wish they taught that stuff everywhere, and earlier. Like for someone who doesn't go to college—I think a really good life is to become an electrician or plumber, for example. They make a lot of money, it's stable, but maybe they didn't get an advanced critical thinking education that would benefit them just living in the world. Why didn't they get that? Why shouldn't they get that?

Hmm, I could probably use a logic class myself. Are there more women in statistics now?

Oh yeah, now it's like half.

Oh really!

Maybe not half? But way more than before. When I was in grad school, there was at least one year when I was really the only female in any of my classes. And even at Penn State, which was my first academic appointment, I was the only female in the department.

I think pretty frequently about how I'm lucky to have grown up directly benefiting from the work that your generation's women have accomplished. I  have never questioned my right or ability to do whatever I want. But then, you still get T-shirts for girl babies that say "Too pretty to do math" and all that, and so many industries like tech are still deeply unfriendly to women. You were part of a vanguard that flipped things so quickly that a statistics department could go, in 35 years, from no women to almost-equal. Do you have any advice?

I actually think what you said earlier, about just working, is really accurate. I mean, where do you get your feeling of self-worth? Is it externally derived or internally derived? I'm not against externally derived self worth—you want people to like your work, there is always some element of that. But really, you have to feel good about yourself. You're the person you need to make happy, not everyone else. Thinking about what society thinks is acceptable behavior, that's fine, but you can't let that stop you. You have to think of #1, and #1 is the person inside, and by making yourself happy you'll actually help society in the end.

And that game I was talking about earlier, it's especially important for women. Women are often asked to play a different sort of game. Understanding that it's a game and understanding that you can play the game consciously without necessarily having to buy into the things it's telling you, makes all the difference. When you do that, that's where you find the places where you really can spend your effort pushing back on it and saying, "Okay, now we're going to do it my way."



15 Comments / Post A Comment

pissy elliott

This interview is the coolest. Loved reading it.

Lucienne

I like, by the way, that this interview talks about some of the promising things technology can do for people.

JuiceBox

Refreshing and inspiring!

RNL
RNL

This was so great! But I still have absolutely no real idea of what she's talking about when she talks about solving abstract problems. No data. Alpha beta gamma. What? It's neat. My mind obviously works in very very different ways.

j-i-a

@RNL hahaha oh my god I have no idea either! "it took me seven months." when she said that, my brain fell out

RNL
RNL

@j-i-a thank goddess for the humanities and finding places where being acutely painfully aware of social cues is an asset!

The Hyperbolic Julia Set

@RNL Hey that's ok! I was a math major and it didn't make sense to me until I was actually doing those problems. Ironically I think the abstract problems can be a lot simpler and better for younger groups (like she said about critical/logical thinking). But they're hard to explain...I guess the first abstract problem I had was like this: They gave us 4 made up rules about how lines could interact (like every line has to touch 2 other lines, etc.) and then they said draw 4 lines such that the picture meets all the criteria. Tada! That's an axiomatic system! You can do geometry proofs about that picture!

EM87

This interview is great! Go Susan Murphy! Also we really should learn logic and stats in high schools; as someone who works in research it's consistently infuriating to me how sloppily research and data are interpreted or presented by media.

crocuta

This was so great!

stuffisthings

I had never thought about minorities being attracted to quantitative fields, but now it makes sense! This helps to explain the many women involved in early computing, most of whom had come from a pure math background. I bet they also saw computers as one of those fields where you could "prove" you worth, before computers become a big-money industry and they got pushed out by tech bros.

I wonder if this also explains why many stigmatized groups (Jews, certain immigrant communities) stereotypically push their kids towards the licensed professions: the idea that once you have that credential, no one can take away your social status and ability to make a good living?

sashay

@stuffisthings Those are some interesting ideas!

The irony is, I think the credentialism works in both ways. As you said, there's an advantage of going into a field where once you have the MD or the CPA or the JD, no one can take that away from you. On the other hand, when a field is so new that there aren't established credentials for it, it can actually be *more* open, because deciding who can be awarded those credentials is a really effective way to shut people out.

I think that's part of the women in computing story--for the most part, women weren't even being admitted to grad school in mathematics, robbing them of a chance to get a PhD.
In contrast, Computer Science basically didn't exist as a separate academic field 'til the 1970s. Necessarily, then, in the early days, being a success in computing had to be about proving your chops by solving big problems, not about your academic pedigree.

I also think there's something to be said for the idea that newer fields are *always* more open to women. In the case of computing, the smart math white men already had plenty of opportunities to do interesting work, but the smart math white women did not. Because the men weren't as interested in the field, there were fewer barriers to entry. No one cared enough to bother stopping the women.
I also think wealthy white women had an advantage over men and women of color, in that it's always been socially acceptable for white women of a certain class to pursue intellectual hobbies. While computing was considered merely an intellectual hobby, they were free to pursue it at will. It's a complicated story.

j-i-a

@apolsasam I DON'T UNDERSTAND YOU

whitty199

Oh boy, oh boy. I've been a Hairpin stalker for over a year, and just created an account to say THANK YOU for this interview. As a female statistician, this is so, so awesome.

The Hyperbolic Julia Set

@whitty199 Yeah! Seconded. I branched off from math into CS but as a STEM lady, thank you for this post!

Post a Comment

You must be logged-in to post a comment.

Login To Your Account