Managing expectations

Noted health economist Dr. Emily Oster talks about self-awareness, resilience and fruit flies

Brown University’s Dr. Emily Oster was Marquette Business’ 2015 Marburg Memorial Lecture keynote speaker

One of the nation’s most respected social scientists began her life as the subject of a social science experiment. At the request of a colleague, Emily Oster’s parents — both Yale economists — recorded the toddler’s earliest ramblings, which were reportedly remarkable for an 18-month-old. The tapes became the subject of Narratives from the Crib, a child psychology and linguistics text.

Today, Oster is a parent herself and the author of Expecting Better, a provocative look inside the statistics behind common prenatal advice.

The celebrated young economist was this year’s Marburg Memorial Lecture speaker. Prior to her address, Oster sat down with Marquette Business to talk about her career, from diapers to dissertation.

Interview by Christopher Stolarski
Edited for clarity and brevity

Both your parents are economists and you practically began life as a research subject. So, talk to me a little about Narratives from the Crib.

Okay, just to be clear, Narratives from the Crib is from when I was an infant. (Laughs) My parents are both economists at Yale, and at some point when I was very young, maybe 18 to 20 months, there was another economist named Dick Nelson who was visiting Yale. Dick’s wife, who is a child psychologist named Katherine Nelson, interacted with my parents at a cocktail party. Katherine was studying early child language development, and my parents mentioned that I had this habit of talking to myself in the crib. Because she was a researcher, Katherine asked my parents if they could record me because researchers didn’t have much chronological evidence on how kids develop languages. So they put these tape recorders under my bed and recorded what I would say for several years.

So now there’s this book, which is a collection of people’s studies that are not so much about what I was saying — which in my sense was quite boring — but about the way that language developed in a particular case study. Example: When do you start hearing me talk about past versus future? Which parts of speech come up first? It’s very much a linguistics book.

So, it’s not beach reading?

It’s not beach reading, that’s for sure. (Laughs) Even if you’re the subject.

Thinking back — and maybe at the time you didn’t realize it — but how do you think some of these early experiences shaped you?

I have always been interested in doing research and the idea of exploring ideas, as well as learning things about the world. And those are still the parts of research I like — those moments when you’re the only person who knows something. When you had a question and you find out an answer and you’re the only person who knows the answer to that question. That’s a very exciting moment.

I think I probably thought about these things earlier than most people, because I had these academic parents and I understood that this is a job that you could have.

Did you ever have a moment of rebellion at any point where you just said, “To heck with social sciences. I’m going all in on art history!” or something?

Oh definitely. Not art history; I’ve always been pretty practical. But I thought I was going to be a biochemistry or medical sciences researcher. Then I went to college.

I was going to major in biochemistry, and the summer after my freshman year I worked in a fruit fly lab; I had two jobs. One was to incinerate dead fruit flies. Fruit flies die pretty quickly, so you transfer the live ones to new test tubes; then you heat the old tubes until the dead ones melt. My second job was to dissect the larvae brains under a microscope. I got terrible RSI [repetitive strain injury]. I just hated it. It wasn’t a job anyone would enjoy, but the deeper issue was that I realized it was going to be a very long time before I was going to actually be able to ask or answer any really big questions about the world. And that did not work for me. So I switched to economics and didn’t look back.

In 2007, the New York Times’ David Leonhardt billed you as one of the most promising young economists, alongside other under-30 scholars like former Marburg Lecturer and Milwaukee native Raj Chetty. Reflecting on that point in your career, talk a little about what it meant to be a rising star among your peers. Did that even matter to you?

I don’t think that was very good [for me]. I think one of the things I wish had gone differently — I mean obviously things have turned out fine — but I got a lot of attention for stuff that I did early on. Now part of it was stuff that was wrong, so obviously I wish I hadn’t been wrong. But I think a lot of people thought I was really good early on, and when that happens it is hard to live up to.

The truth is, if I look back at that list, most of the people on there are really, really successful economists. I am pretty successful, but I am no Raj Chetty or Jesse Shapiro or Matt Gentzskow. That was a hard thing to live up to. I think at the time I probably thought it was really cool, as opposed to now that I’m more of an adult.

Oster with Dr. Andrew Hanson, Marquette Business economist. Hanson’s work on signaling theory in sports economics was featured in 2014.

In that same piece, Leonhardt wrote: “Whatever you think of [Ms. Oster’s] conclusions, there’s no denying that her subject is more interesting — and, yes, more important — than the esoteric fiscal and monetary models that once dominated economics. Ms. Oster is studying death, not taxes.” What he’s saying, I think, is that your work and the work of those in your cohort, is more accessible — more pedestrian friendly — than the economic studies of past generations. Why that shift? Why now?

The big change is in how people are approaching data, and how accessible they’re trying to make their conclusions. I teach a class on empirical methods, and we think about the evolution of empirical methods over time — there’s a lot of push in economics towards the credibility of data and credibility of conclusions. We try to think about questions in the world and answer them in ways that will be convincing to other people. The data has improved, but the methods, as a result of the better data, have gotten easier to explain. That has pushed our influence externally because people can understand what it is you’re saying and why you’re drawing your conclusions.

Oster cites Steven Levitt, author of best-selling ‘Freakonomics,’ as one of her — and economics’ — key influences. Oster is featured in his follow-up ‘SuperFreakonomics.’

We can probably trace this all to a few different people, particularly Steven Levitt. For sure, [his first book] Freakonomics was a big deal in making people think about economics as something other than taxes — or at least easier to explain to the guy sitting next to me on the airplane who wants information about his stock portfolio. Those kinds of interactions have faded over time as people have realized that economics does something else, and I think it would be a mistake to undervalue Steve’s role in that. Steve is also someone that I am very close to, and he had a very big influence on the way I think about doing research. The decision to write a popular book was something that I talked to him a lot about.

Your dissertation on Hepatitis B and the “missing women” garnered a lot of attention. Three years later, you wrote a follow-up paper that essentially stated your original findings were wrong. This sort of mea culpa takes a keen level of self-awareness — why is that important in your work?

Just to be clear, at the time this was the worst. Up until not getting tenure, this was by far the worst professional experience I’ve ever had. Sometimes the way it’s told, it comes across like, “Isn’t it great that you admitted you were wrong?” That definitely isn’t how it was at the time. It was a hard thing to do.

There had been a lot of initial push-back on that paper and it pushed me to get new data. Then there was a point at which I was at some conference, and I pulled up this data on my computer and it was obvious from the regression that it just wasn’t there. I had hoped that it would be, and I wanted to see it — it just was not there.

I remember the next day talking to one of my colleagues saying, “I ran this analysis check, I got this data and it’s just not there. I am just going to have to write something that says it’s just not there.” He was like, “Okay.”

It’s that moment when you realize there really isn’t another path. You always hope when these sorts of things come up about your work…where you hope there’s a way out. Eventually it was just, Nope, not there! And there’s really nothing I could do. It was better to just say it’s not there and let’s move on.

You said at the time that it was not a great situation. But I think eventually you were sort of lauded for your resiliency. How did that impact how you treated future research projects?

I was more careful. (Laughs) The thing I had wished I had done with that project — obviously ex-post — is to do this other analysis before I wrote the paper in the first place. And I think I have tried with subsequent work to push more on what else can I do to convince myself that this is right, or try to convince myself that this is not right. And I think that that’s the biggest legacy.

In 2007, Oster delivered a well-received Ted Talk on AIDS in Africa.

Fast-forwarding a little, you now have the book Expecting Better, which received some mixed reviews. Some were very good. The New York Times had one that was a little less good. What do you say to that?

The main thing is don’t read Internet comments! (Laughs) That is a good life lesson, which I highly recommend. You know, look, I think with something like the book it’s easier to take those kinds of criticisms in a more accepting way because that’s the nature of writing something for the popular press. Some people are going to like it and some people aren’t going to like it. Obviously, when I read the New York Times review I felt like [I wanted] to be able to come back and say, “But you’re not right.” She had some stuff about [drinking] alcohol [during pregnancy] and then she wrote, “Everybody knows all of these things.” And I thought, “No! Everybody doesn’t know all of these things!” Of course, you don’t get to argue with the New York Times.

It is different than an academic setting because people are coming from a lot of different places, and some people aren’t going to like your tone and they just don’t like your personality and that comes out much more. Academic work does not have a personality and the book does.

In the end, the long-term response has been very different than the initial response. A lot of the initial discussion was about alcohol. I get two emails a week from women with questions from the book. No one has ever written me a question about alcohol. All of the questions are like, “I’m 41 weeks and two days pregnant, and I’m worried about this aspect of the fluid and is it enough to stay hydrated?”

I suspect for many expectant mothers, a mere list of rules is welcome, almost comforting. People in stressful situations often want simplicity. Do x, don’t do y. With respect to pregnancy, is there any danger in that?

I’m not sure. If people want that kind of advice, I think that’s fine. The danger is when you start thinking everybody wants that advice. Or worse than that, when you start believing that advice has actual content. For instance, there’s this ‘rule’ that 37 weeks is full term. There is nothing special about 37 weeks and one day of gestation versus 36 weeks and six days of gestation. It’s one more day. But at some point there was some definition that 37 weeks counts as full term. And somebody came up with a good rule of thumb and it’s a reasonable thing. But over time, people started treating that as if it had content.

So, basically, it was like, “You’re after 37 weeks so we can start thinking about inducing labor because all babies after 37 weeks do the same.” There have actually been a fairly large number of early term inductions where they induce labor significantly before the due date with this idea that somehow 37 weeks was a meaningful number. And in the end they’ve now had to come up with — since that turned out to not be true — they’ve had to come up with additional guidelines which say don’t do that. You have to wait until 39 weeks.

That kind of reaction to these rules is one problem with a rule. It’s fine to say, “We’re going to have this rule of thumb,” and if people want a yes/no guideline we can give it to them. But when you start thinking that the rule has content or that it’s some sort of magical rule from the sky, that is the problem.

Guideline versus gospel.


So, lastly, what’s next? I hear you have a working paper on diabetes.

This is a paper about trying to estimate how people respond to health news, which is what a lot of my research is about at the moment. The particular question is how one makes household dietary choices in response to a diabetes diagnosis. The innovation in the paper is to use some household-based scanner data, where households are asked to scan all of the items that they purchase at grocery stores and drugstores, so I can see at the UPC level what you buy. I use purchases of glucose-testing products after some exclusion period as a marker for diagnosis.

So, we see people who are not purchasing testing strips and at some point they start purchasing testing strips. That tells us that they learned that somebody has diabetes. And then we look at what happens to their calories purchased after that. It’s a setting in which we can see a lot about how much they change and in what particular ways…what kinds of people are more responsive or less responsive.

Watch Oster deliver the 2015 Marburg Memorial Lecture in its entirety online.