17. Emily Oster: “I Am a Woman Who Is Prominently Discussing Vaginas.”People I (Mostly) Admire
- 1,520 views
- 27 Feb 2021
In addition to publishing best-selling books about pregnancy and child-rearing, Emily Oster is a respected economist at Brown University. Over the course of the pandemic, she’s become the primary collector of data about Covid-19 in schools. Steve and Emily discuss how she became an advocate for school reopening, how economists think differently from the average person, and whether pregnant women really need to avoid coffee.
Emily Oster is not only a leading academic economist tenured at Brown University, but also through her popular writing, one of the most trusted sources on the subject of pregnancy and childbirth. And over the last year, she's been an early and influential voice advocating for school reopening. I often say economics is just the thoughtful application of common sense. And Emily, more than just about anyone else I know, has found ways to take what is useful about economics and integrated into all aspects of everyday life.
And she's been doing that since a very tender age. When she was only nine years old, she started a weekly newsletter for the residents of her block, replete with data analysis, pie charts and graphs. And even before that, as a precocious two year old, her nighttime ramblings lying alone in bed were so remarkable they were tape recorded and analyzed, resulting in a book called Narratives from the Crib, which is still in print today more than 30 years later.
Now with an introduction by Emily Oster herself. Welcome to people I mostly admire with Steve Levitt. I first met Emily Oster almost 20 years ago, I was visiting the Harvard Economics Department for a day to give a lecture to the economics faculty.
And as is customary with such visits, the remainder of your day is filled in with one on one meetings with faculty members, and sometimes they fill in the slots with the most promising graduate students. As I looked over my schedule that day, there was only one name I didn't recognize. Who is this Emily Oster? I asked.
One of your star PhD candidates. The organizer responded. Well, actually, she's one of our undergraduates, but I think you'll find it worth your time. And I have to say, I was deeply perplexed until I met Emily and she began to describe her undergraduate thesis on the relationship between crop failures in the Middle Ages and the frequency of witch trials. It was like I was talking to a young version of myself, except that she was female and 10 times better than I was at her age, doing the weird sort of economics I always loved.
I've been her biggest fan ever since. And you know what's so funny about the whole thing?
I remember meeting like it was yesterday, every detail. But Emily has no recollection of it at all.
It's supposed to be the other way around when the hotshot professor meets with the undergraduate.
Emily Oster, I'm so glad we're having the chance to talk today. I am really excited to get to talk to you. Thanks for having me.
So as you well know, I think the academic research that you've done is mind blowing. But I'd like to start with your transition from being an academic economist to being an all purpose purveyor of thoughtful, well-informed, commonsense.
And I think that story begins when you were pregnant with your daughter and trying to make good choices. Yeah, I think that's roughly right. I've always been pretty interested in trying to communicate, research and communicate the ideas of economics to the outside world, but I got much more interested in it. When I got pregnant, I started thinking about using things that I was doing in my job, in my pregnancy, and just trying to figure out what the data really said.
And I got obsessed, maybe a little strong, but yeah, like a little bit obsessive.
And you were disappointed in the advice that was available to expectant mothers. I found the way that many of the pregnancy books were written and much of the advice that I was given to be basically pretty patronizing, it was in a space of either just don't do anything and don't think about it for yourself or or everything is going to be fine. Little girl like Pat Pat on the head. Just listen to the doctors that didn't work for me.
And so you decided in the midst of all your academic success, but still young and untenured to do something that many people would perceive as crazy and that you decided to write a book for a nonacademic audience on pregnancy, even though you're an economist. And that book, of course, was called Expecting Better. Did you not think that was career suicide?
I basically didn't think that much about it. I had been doing some of this research and doing some writing, and I really liked doing the writing. I thought it was really fun. And on a total whim, I sent a book proposal to Susanne Gluck, who is your agent and is now my agent and who's been on this podcast, by the way.
And I sent it to her and I said, you know, I'm working on this issue, wrote back and was like, this is great. Well, I have a few comments. We'll send it out. And now that I know more about publishing, I realize how bizarre and lucky that was. But I thought at the time that I could pitch it as like, it's a hobby. I'm also doing my academic work and this is a thing I do on the side, like windsurfing or something.
You're not allowed to have hobbies when you're an academic. I mean, you have to pretend you don't have hobbies, even though we all do. So just let me describe the book for people. You come at it admittedly as an outsider and you don't pretend to have any particular knowledge. You just say, hey, I'm going to look for the research that meets basic thresholds and I'm going to look at those as an educated data scientist and tell you a woman what they say and then you get to make your own decisions, given the evidence and what the costs and the benefits are.
I think the strength of the book to me was that you come off as being very believable because you don't seem to have any agenda at all.
Yeah, I think the agenda was to try to figure out what was the good research. I mean, sometimes people ask, what's a book like? And I say, well, it's something between a memoir and a meta analysis. There's a piece that's about my personal journey and this. But then a lot of the book is like, here are a bunch of papers on this. Here's how you can wade through this literature and understand what it really says.
And you're capable of really looking at this data and making these decisions for yourself. And you don't have to trust me as some expert, except that I'm an expert in telling you what the data says, but I'm not an expert in what you should do.
So could you just give us one example, one of the topics so women are told to restrict their coffee intake or not have any coffee? The advice varies and the idea is that it could increase the risk of miscarriage. I ask the question both. What is the data really say? And also why do people disagree, which I think is important for people to understand. If you compare women who drink coffee to women who don't at some pretty high level of coffee drinking, you will see an observational link with miscarriage.
But it is also true that the women who drink coffee are very different on a bunch of other dimensions, most notably age, which is also correlated with miscarriage. And it's hard to know if it's age or coffee drinking. Coffee drinking is also negatively correlated with nausea, so women who are nauseous tend to drink less coffee. But nausea is also associated with a decreased risk of miscarriage. So as you add controls for age, for nausea, you get weaker and weaker correlations trending towards zero and in many cases actually being zero.
So let me just make sure I understand if you just look at raw data, not trying to take anything else into account, what you see is that there is a positive relationship between caffeine consumption and miscarriage. And then as people are more careful now to say, oh, well, older women drink more coffee and they have more miscarriages, let's try to compare older women to older women and different variations on that. What you're saying is that the correlation you see in the data changes to become weaker and weaker.
So from that, you conclude that probably the sensible thing is to drink, maybe not infinite amount of coffee, but not to worry about it very much. That is correct.
And so that's just an example of logic. And you do that for every topic and you do it in a way that brings a lot of reassurance to people. And it's interesting because I think part of the reassurance comes because you're an outsider and you haven't done any of this research, you don't have any stake in it. And so you feel like an honest purveyor of the truth.
I actually have come over time to also think the fact that I was doing this in the service of my own pregnancy and that I talk through how I thought about some of these choices, has made the book more accessible to people and made them more comfortable because they feel like I'm their friend, which I am happy to be anybody's friend.
So cryptid, which is your second book about child rearing, is very much in the same spirit as expecting better. But in many ways I think. Harder to write because the data are not nearly as extensive or as conclusive on childrearing as they are on pregnancy. So how did you do with that?
When I came into writing crib sheet, I thought it would be much more like expecting better where it would be just grinding through a lot of papers and there's a piece of that. But then it also becomes clear as you do this, that actually a lot of what you conclude at the end of extensively researching a topic in early parenting probably doesn't matter too much in the long term. And that meant that in the end, much of the message of crib sheet is actually there's a lot of good ways to parent and you really need to pick the things that are going to work for you and your family.
Can we talk about breastfeeding?
Because I think that's such an interesting debate as my wife was facing the choice of breastfeeding or not, she had been completely and utterly convinced by the public discussion that she would be an awful, awful parent if breastfeeding didn't work on. The public pressure that people are putting on themselves, on each other about breastfeeding is, as you say, it's so far beyond anything that could even be supported by science.
I read some article at some point that said breastfeeding will help you form better friendships.
What would that even mean? It's become such a cult thing. And I think that the downside of that is that if it doesn't work or it's not for you or you don't like it or you want to quit, it's like, well, don't you love your baby? And then people feel terrible.
Yeah, absolutely. So just for the record, let's just talk about what the science says about breastfeeding, because there's this mythology about it. And then there's the actual facts.
When you dig into the science, you find that breastfeeding has some benefits to digestion early on in life. So particularly for babies who were a few weeks old in the first year, maybe even there are some lower gastrointestinal risks, maybe a slightly lower risk of ear infection that's a little weaker. There are some also digestive benefits for very preterm babies. And there's actually some evidence of a breast cancer risk reduction for the mom, not for the baby, but all of the kinds of things that your wife is hearing about and I heard about around your kid will be smarter and thinner and they'll be able to fly and they're going to have all these long term life health benefits.
Those things are just not supported in the data.
And how about IQ? What's the evidence on IQ now? Not IQ zero zero zero.
So you're happily going along an academic who's writing these popular books, and now it's March 20 20, what happens because you've been writing these pieces on what if your kid can't poop and you really suddenly change course?
My publisher said you should really have a newsletter. I started in January 20 20 and I had this image that this was going to be just an occasional thing to connect with the readers. And then when covid started, all of the questions I got were just things about covid and critical of it. And I started writing a lot about covid both around some of the pieces of data on how risky is it for your kids and then also much more on decision making and trying to help people think through some of these decisions that they had never thought that they would need to make, particularly around parenting.
You know, should I see the grandparents? Should I send my kid to daycare? And then it sort of evolved into studying daycares and schools.
So I think you've developed a framework for making sensible decisions in settings in which you don't have great information to describe a little bit of how you think about.
Many economists talk a lot about trading off risks and benefits, which is a big piece of decision making. But when I outline this for people, I made the point you need to start. By framing the question you are asking, and so particularly when there's so much uncertainty and the decision is so new, I think that people very frequently, they know what the first option is and then they frame the second option as or not, should I send my kid to daycare or not, but or not is actually not an option or not as some other thing.
And this choice is going to be really different if it's should I send my kid to daycare or hire full time that daddy, should I send my kid to daycare or have my elderly parents take care of them? Should I send my kids to daycare or quit my job? And then I tell people, look, you need to end your decision by making a final decision rather than what I think a lot of people do, which is just keep thinking about it over and over again and allowing it to take over every time they're in the shower, every five minute conversation with their spouse, every little moment.
You really want to say, OK, we're going to make the decision and then we're going to try to move on.
So you have this decision framework, but somehow or another, you went from that into being a creator of data.
How did that happen?
I mean, who knows? I really like data. And I started to get just very frustrated at the lack of data around kids and particularly around kids and child care. And like to be totally frank, my children were at my house all the time and I was eager to dispense with them to an outside location. I just put up a Google form in my newsletter. I had by the end of it, like one hundred thousand kids. And the corporate rates were like about a tenth of one percent, which I think turned out to be very, very close to what other people ultimately found in better data.
And you go from there to what I think is a really unlikely step of becoming the central collector and keeper of data on kids in school and covid for the entire U.S.. I mean, God bless you for doing it, but how did that happen as a result of this child care stuff?
I started writing more about schools and what it would take to open schools. And then at some point, somebody connected me to a data processing company and to the School Superintendents Association. In the end of September, we published some data on like two hundred thousand in-person students, a very, very small number of kids. But then, of course, nobody had any data uncovered cases in schools. And then some states started doing some things that were useful in helping us a little bit.
And so then we became this aggregator source.
I should have said this is called the covid School Dashboard. Yes. So what have you found?
These are the rates in schools are low. They mimic community rates. But I think relative to some of the theories that people had in the summer, that schools were going to be super spreaders and giant covid sources, that isn't showing up in the data.
And is there a difference between the positive rates for the students and the teachers? Definitely.
So you see higher rates in teachers and staff than and students. And, you know, that is almost certainly reflecting, at least in large part, the fact that staff are older and in general, people who are older tend to get covered at higher rates than people who are younger. So kids just have very low rates. So I have to admit, I find the results surprising because I would just think kids aren't necessary that great at keeping their Mahsun in the news a lot, I would have expected more spreading to you or anyone else.
Have a good idea why schools aren't spreading as much as we might have thought.
I think your instinct is insane. Many of us had earlier. I think it's been a hard thing to push out of our minds because kids are generally pretty high risk for viruses like this. But for whatever reason, kids are much, much less likely to get covid, particularly little kids. I think the other thing is that a lot of schools that have been open are pretty controlled environments. People are wearing masks. They're anxious in school, which is, of course, in some ways bad, but also means that they're being a little bit more careful than they would be otherwise, particularly the adults.
On the other side of the coin is that in deciding whether to reopen schools, it depends on exactly how terrible remote schooling turns out to be. What does the evidence look like in that dimension?
Remote schooling has been really terrible for people. A lot of kids are just not logging on. And so you can't teach kids if they are not there. Another thing is that it's just hard for little kids in particular to focus on Zoome. The other thing is the kids are really struggling emotionally being out of school. And so we've seen learning losses, we've seen increased failure rates. We've seen more reports of anxiety and depression and weight gain and all of those auxiliary add on effects.
So the way to the evidence, based on what you're seeing of relatively low transmission rates, big costs of remote schooling, have led you to transition away from being a data collected to really being an advocate for reopening schools. Is that a fair characterization of your position?
Yeah, I think that's fair. I think I came into this conversation really wanting to see what the data said, understanding that there were likely to be learning losses. I think nobody anticipated how large it would be and how does it feel? I hate it.
I am doing it because I think it is important for kids and because I think it matters.
So to put it mildly, not everyone has been such a fan of your no that efforts now. But let me ask you a question. People can be unhappy with you because there's disagreements about what the data say. So one person's status is one thing and one person's data says another thing. Or everyone can agree on what the data say, but differ in the interpretation of the data and this particular issue. Do you have a sense of whether the disagreement is about the data or about the interpretation or bubble?
I think it's a little bit of both. I think early on people were saying, look, this is an entirely opt in sample and that was initially true. That's definitely no longer true. And then I think that, you know, the conversation shifted a little bit to, well, how can we be sure that there isn't some additional risk to staff? And you can't be sure about that from our data, probably from any data. And I think that there's some difference in what we should conclude based on what we see in the data.
So if we thought that being a teacher had a similar rest of the population or a slightly higher risk, like, how should we think about decision making there?
And I think that is at peace where there was some disagreement, I almost wonder whether it be a partial solution isn't something like combat pay, where just like we're soldiers because of some extra risk that we're putting on teachers, maybe we should be compensating them for the risk.
That's certainly something that people raise that at the beginning. I mean, I think we also want to be clear that, like, the added risk of teaching in person in a school is probably substantially less than the added risk of working in a grocery store or driving Uber or working in a restaurant. There's a reasonable view that all essential work should effectively have combat pay in this environment. And I think that you could view some of the prioritization of teachers in vaccines as a version of that thing.
I think they should be the front of the vaccine line. Teachers are an easy group to reach. We know where to find them. They tend to be very socially engaged. If they are vaccinated, I think they will be great advocates for vaccination in particularly underserved communities. I think that because they're out in the world, they are at more risk than someone who is literally not out at all. And also, I think that in some places, teacher vaccination is going to be a prerequisite to get kids back in school.
So I've tried to read the things that are critical of you in this area and. It seems like maybe the worst critique that's been leveled at you is people claim that you downplay evidence that doesn't support your position, which is actually what advocates do all the time. You have to have a simple message and you can't offer the nuanced messages that you've always been putting forth before. It's more of an outsider looking in. Do you feel any conflict in being an advocate?
Interesting question. I think there is a piece of that in advocacy where you want to be in some ways more critical of the pieces that disagree with you versus the pieces that agree. And, you know, I've tried to take off my advocacy hat when I'm thinking about things that disagree.
So obviously we're operating in a really data poor environment when we're making decisions about covid, even despite of your own efforts to try to increase the flow of data. And I wonder whether in your own books, the kind of evidence we have about covid wouldn't meet the standards that you would hold studies up to make these hard choices. Now, I understand we have to make choices here. Do you have any misgivings in light of that?
Not really, because of the last thing that you said, which is that I think that we have to make a choice. People have been thinking about it like, well, let's just be safe and not open schools. And until the data is perfect and we're 100 percent sure that we know exactly how to make sure there's no transmission in schools, let's just keep them closed. But the thing is that we cannot afford to make decisions like that in this current environment because every day that kids are not in school, they are losing.
And so there's no safe, great option here. We're going to have to make a choice with imperfect data. For me, that's been a hallmark of all of the aspects of the pandemic, is we're making all of these choices with not very good data and we're just trying to make the best choices that we can and the choices that minimize harm. You're listening to people I mostly admire with Steve Levitt and his conversation with economist Emily Oster, after this short break, they'll return to talk about Emily's academic studies of health fads and genetic testing.
So I have to say, it blows my mind that if Emily Oster hadn't taken the initiative to start collecting data on Cobden schools, there simply would have been no information on the topic last fall. How could it be that no governmental agency and none of the national organizations of teachers or schools thought to do it?
It is just so symptomatic of our botched national response to covid to my plan for the rest of our conversation is the first dig into some of the most interesting academic papers and then to see if she's willing to talk to me about her one huge academic misstep and why she thinks there's so much animosity towards her on the part of other economists.
You've had so many interesting academic research findings. Maybe we could start with one of your newer studies which demonstrates the self-fulfilling prophecy of health fat. Could you explain that?
I think it's easiest to explain this by starting with just an example. Let's say that some researchers are researching what kinds of foods improve longevity. And they brought a bunch of things and they come up with pineapple's pineapple's, increase your longevity. And let's just, like, imagine that's a false positive, right? They make a mistake. They make a mistake.
Yeah, but they get a New York Times and Washington Post headline that says pineapples are the secret to a long life. Exactly.
And so then you think about what happens. Maybe on average, everybody eats a little more pineapple. But it's particularly the people who are like reading The New York Times to find out what food they should eat to make themselves live longer. It's like my dad. This paper is totally inspired by my dad, who adopts every weird food health fad that The New York Times says that he should do. And of course, those people are also doing all kinds of other stuff, they're also not smoking and they tend to be better educated and they're running all the time.
And so then if you said, OK, let me go back and let me take a look at this pineapple thing again with more new data. Now, actually, you've created a situation where pineapples are even more strongly associated with living a long time because of the people who have started eating them just to make sure that people understand.
So the first go round, the researchers made the mistake. They just looked at a bunch of people and they happened to see the people who ate more pineapple happen to live longer. But we've posited that was just by accident. But then what happens is people like your dad now become daily consumers of pineapple. And so they look three years later and say, well, let's see if our theory still true. But now when you compare the people who eat pineapple to the people who don't eat pineapple, pineapple eaters are super healthy, exercising, good eating, probiotic taking people and all of the Cheetos eating sort of drinking people, they haven't paid any attention.
And so now the comparison is between a really healthy group of people and a really unhealthy group of people. And it's a magnified any possible direct causal effect of the health intervention itself. Yeah, exactly.
As you said at the beginning, a self-fulfilling prophecy where when I tell people they should do something, then the selection of who does it changes. And then you get potentially these magnified effects. And then the question was just can you see that showing up in data and would it affect what we conclude? And so in this paper, I look at a few things, the most striking of which is vitamin E, where there was a period in the early 90s when people were told, like eating vitamin E helps you live for a long time.
And then at some point there was a new set of studies that were like, oh, just kidding. It kills you is actually the second set where randomized studies. So there was some observational studies that said vitamin D is good for you. And then there was a randomized studies that were actually too much of it will kill you. And so what you see is after the studies that say it's good for you, a bunch of people start taking it.
But they're precisely better educated people who don't smoke, who exercise more, who eat a good diet, and then those people just adopt later. But what's really striking is you see in some of the data, for example, the links that you would estimate between vitamin E and mortality basically are created by this election. So in this intermediate period of time when vitamin E is highly recommended, it looks like it prevents you from dying. But that's entirely because of the adoption patterns that we see after we make the recommendation.
I love that because to me, the best research is exactly this, where you take on a topic and the insight you come up with is completely obvious. And I understand it immediately and it makes sense to me. And I find it in my brain and I use it for after. And I ask myself, why didn't I think of that? How could that never have been mentioned before? Those are hard to come up with, but you make it look easy.
How do you come up with ideas like that?
So that idea really came out of the books because I started thinking about like, well, I'm telling people to do stuff and I wonder if they'll do stuff. And at some point somebody has this crazy question that was like, are you worried that all babies will now be the same because everyone is doing the things that you tell them?
It's like, no, I'm not worried about that.
Well, I think what's interesting in that is oftentimes these really good ideas are just hanging in the air. Right? So someone brought that up and you could have just dismissed it. But you saw the kernel of what was obvious and subtle in it. That's the secret to my own research, is I just live my life with my eyes open, always looking for things that are obvious but can't be seen.
And when I very rarely let anyone, I really hold on tight and I try and prove it.
For me, that's like the best moment of doing research is when you have a picture and the picture is what you thought it would be. That's a cool moment.
You've also spent years researching the decisions people make around genetic testing when they're at risk for having dreadful genetic diseases like Huntington's disease. Could you describe that research agenda and what you found?
I got into that by thinking about human capital theory. Human capital theory is how we understand different choices people make about what we call human capital, which would be things like their education or their job training choices, which is a very old idea in economics and trying to think about whether people would invest less in their human capital, get less education, do less job training and things like that. If they knew that, they would live a shorter amount of time.
And so the sort of insight in this set of projects was to look at people who are at risk for this particular genetic disease, Huntington's disease, where the genetics is very simple, if you have one. Copy of the affected gene, you get the disease for sure, there's a genetic lottery, which is if you have a parent with the disease is a 50 50 chance, you get it.
So another key about Huntington's disease is that it doesn't manifest itself until later in life when it appears varies based on some aspects of the genes.
The most common onset time would be the 30s or even 40s, although it can have onset in childhood or it can be very late.
So you were able to see choices people were making as a function of what they knew would happen later in their life.
If they carried the gene, then they would expect a much shorter life expectancy and you could see what investments they were making in school, on the job training, etc.. That was a starting point for your research. But actually, in the end, the thing that turned out to me to be much more interesting and much more challenging to economists was the fact that. Very few people who have Huntington's disease ever get tested, which doesn't fit well with simple economic theory.
Before they had this test, they asked people, are you going to be interested in having this test? And people basically said, yes, like a lot of people said yes. And it wants the test is available. Nobody wanted it. It's a perfectly predictive genetic test that will either tell you're going to have it or not.
And from a simple economic perspective, you think, well, if I know whether I have this gene or not, I'll make completely different investments over my life. Maybe my decision about whether to have kids, maybe whether to go to college, whether to smoke or smoke crack might depend a lot on the outcome of this genetic test. Usually, economists think when there's a test that provides lots of information and you can get it essentially for free, everybody would do it.
But what percentage of the people in your sample actually do get the test is teens so over these people who have a parent with the disease, five percent of them get this test.
It's very small.
A strong economic prediction would be that everybody would get this test or almost everybody, and yet their symptoms are completely wrong. What do you replace your models with whom we started working on this question?
What I wanted to understand was how can we understand this behavior? Why is it that people don't know about the tests, the tests, expensive, something about insurance. And none of those things were consistent with the data. And we actually had some data where we asked people, why don't you want to get this test? And the answer was, I don't want to know. But it started us down the road of thinking, well, maybe there is something deeper to the I don't want to know.
We started thinking about models or mathematical formulations where people get value from what they expect to happen in the future, and in particular, they get value from imagining what's going to happen in the future. The basic idea is like if I knew for sure that I was going to develop this terrible, debilitating disease later in life, I would have a hard time imagining that wasn't true. But if I don't know, then I can basically pretend that it's not a risk that I can have this anticipation of having a healthy, long life and that I may value that anticipation in a concrete way, so much so that I'm willing to not get the test.
I've asked a handful of economists about this, and I've asked far more, not economists.
And what is so interesting is that every economist immediately says, of course, I would get the test. And literally every non economist I've ever asked has said that would never get the test.
And it's completely obvious to them. It seems, at least the way I describe it, that they wouldn't really want to have the test. And there are a handful of other issues where I've done the same thing. I think a market for organs where people could buy and sell their organs. I think almost all economists are in favor of such a thing and almost all non economists are against it.
It's always interesting and surprising to me to see that there is this really sharp distinction between people who are economists and people who are not.
When I explain this to people who are not economists, like, of course nobody would want to know, like, why would you even ever think anybody would want to know this information?
The theory said, so you've had these many wonderful, remarkable papers.
And then you also had one giant misstep where one of your findings ultimately didn't turn out to hold up. Would you be willing to talk about that? This whole thing is your fault.
So I think we're just at the front. Let's just say I know everything we're about to say is one hundred percent, you're gone, OK?
And I'll take lots of responsibility. So I was indeed the editor at the Journal of Political Economy at the time, and I handled your paper and I absolutely will take some fair share of the blame.
But so just describe what the paper tried to do in the beginning.
There was this existing body of literature, most prominently popularized by Amartya Sen, noting that there are more men than women in a number of Asian countries which are known for gender bias. And this is traditionally attributed to neglect of female children or sex selective abortion in some cases.
And we're talking about a huge number, 100 million missing women. And the basic conclusion at the time was that they're missing either because of sex, selected abortion, or of increased mortality rates along the way because of neglect by parents, essentially.
Yeah, exactly. You had a different theory. I had a different theory based on having read a book and some science data. And my theory was that high rates of hepatitis B in these countries was contributing to some of this gender gap. And the idea was that women with hepatitis B were more likely to have male children, perhaps because of some aspect of interactions in the womb, and that was driving differences in the gender ratio in the particularly in China, where hepatitis B, cancer rates are very high, that that might actually explain principle quite a lot of these missing women like.
Half of them or something, just to be clear, there was scientific evidence that suggested this might be true in your role, really was to take that scientific theory and try to quantify it by going to data that existed. And you did that in a number of very imperfect data sets. But the best that we had and you found really surprising and compelling evidence that it was important. And I found that evidence very convincing and eventually published your paper. And then people started looking at better data.
As you said, the sort of data sources that I was using, I thought were compelling but were imperfect. And then at some point after the paper was published, some people got access to really excellent data on this, which had actual testing results, and they found basically no effect of being a hepatitis B carrier on the child gender.
I remember asking you, I said, what would be the definitive data source? And you basically designed something which would be the definitive study, knowing that most likely it was probably going to go against you because the better data had supported the other side of the story.
And I really admired the fact that you went and invested years of your life to go and find the truth, even knowing going in the truth was probably going to work against your initial hypothesis. And in the end it did.
And of course, I hoped that I would turn out to be right. But given what we knew at the time, it did seem likely that what we would find was that we were not. And, you know, I went to China, I collected a ton of data, went around to a lot of places in China and talked to people and got the data from registry books. And then, yes, in the end, I was not right.
And you wrote it up frankly and honestly, and you crushed your original paper in the process. And, well, that is how science is supposed to work. I literally cannot think of another example in economics of that ever happening. You know, what makes me really mad is the fact that our profession should have cheered you on and should have canonized you for being a great scholar. But the only thing people wanted to talk about was how you, quote, really blew it on hepatitis B, unquote.
So I just want to go on record as saying, I think what you did was awesome. And I'm so disappointed that not that many people have recognized that you just did the right thing.
That's nice. Thanks. Things turned out OK and was an important learning experience, I think, for me. You know, I think if I had been a more mature scholar, I probably would have tried to get that data first before I published. Get more cautious as we age maybe.
So within the economics profession, I get the feeling that a lot of people don't like you, just like a lot of people don't like me. And I think that there's a lot of similarities in the way you and I approach the world.
But you also have the feeling a lot of people don't like you.
Oh, definitely. Yeah.
I have to be honest, while in general I'm almost oblivious to most gender based arguments, I have the feeling that you've had a harder time in economics because you are a woman and you really have come out with strong opinions on important questions. And I'm sad to say, I think that you're being a woman is part of why people have had a hard time accepting some of what you've said. Do you have any sense of that being true?
There may be some gender disadvantages. I think there are other ways in which I'm tremendously advantaged by my parents are economists. And so in those ways, I feel very lucky. Economics is a weird it's a very gendered place. There are a lot of women who have really awful stories about what it is like to be a female economist, and I have a few of them, too. I think it's not just that I'm a woman, but I am a woman who is prominently discussing vaginas.
And that's not just in my looks. I mean, I once wrote a paper about menstrual cups in Nepal. A menstrual cup is a reusable cup that you use during menstruation. And when I give talks about that, I would bring a menstrual cup to the talk and pass it around. There's a being a woman and then there's not being apologetic for it. And I think that's not to everybody's taste.
Do you have any advice for people who find themselves in the middle of a big public fight? Try not to take it so personally, that's my personal advice to myself. I think part of what's made this advocacy position, particularly around schools, so unpleasant for me is that I take everything like really, really, really personally. It's like when you read your teaching ratings when everyone's like, oh, you're an amazing teacher. And then the one guy is like, this class sucked, but you're in a public fight.
There's a lot of guys who are like that guy.
What about emerging with your reputation intact or well-liked or respected? You have advice about that? Try to be honest and try to be honest when there are limitations, and that's hard, as we said, move into doing more advocacy, it becomes more difficult to keep that. But I think that at the end of any public fight, and certainly for someone like me, where I hope that I will stop doing this stuff on schools and they will all open and I can go back to, you know, writing about Jews, I want to come out of that, knowing that I feel comfortable about what I did with me and that I'm the person I ultimately have to live with.
And my husband, I guess. Believe it or not, in 25 years, I have never read my own teaching evaluations. I want a teaching award. The very first time I taught, which I took as evidence that I was a naturally born, amazing teacher. And ever since then, I've been terrified to look at my teaching evaluations because what if it's not true?
Well, at one level, I know it's not true.
It's been 25 years since I won a teaching award. And that's pretty solid evidence that I'm not quite the teacher I imagined myself to be.
In the end, when it comes to teacher evaluations, I guess I'm just like the people Emily Oster studied who chose not to take the genetic test to determine whether they carried the gene for Huntington's disease.
Sometimes ignorance is bliss, even when the economic models tell us otherwise.
People I mostly admire is part of the Freakonomics Radio Network, which also includes Freakonomics Radio, No Stupid Questions and coming soon, Sadir breaks the Internet. This show is produced by Freakonomics Radio and Stitcher. Morgan Levy is our producer and Dan DeSilva is the engineer. Our staff also includes Allison Crichlow, Mark McCluskey, Greg Rippin and Amateur. We had help on this episode from James Foster. All of the music you heard on the show was composed by Louis Scarra to listen ad free, subscribe to Stitcher Premium.
We can be reached at Tima at Freakonomics Dotcom. That's P. I am a at Freakonomics Dotcom.
Thanks for listening. Maybe we're jerks. Yeah, we might be jerks. Ditcher.