Online Journal | Aaron Mathew | January 2024
Dr. Katy Milkman is the James G. Dinan Professor at The Wharton School of the University of Pennsylvania and the co-director of the Behavioral Change for Good Initiative alongside Dr. Angela Duckworth. Her research interests span behavioral economics, psychology, and research methodology, which she explores in her NYT Best-Selling book “How to Change: The Science of Getting from Where You Are to Where You Want to Be”. Dr. Milkman also hosts Choiceology, a leading behavioral science podcast with Charles Schwab. I had the pleasure of speaking with Dr. Milkman on January 18th, 2024, about her research origins, the impact of BCFG, and the everyday importance of behavioral economics.
Aaron Mathew: Good afternoon. Thank you so much for taking the time to talk with me, I know you have a really busy schedule so I really appreciate it.
Katy Milkman: Oh yeah, I’m happy to do it.
AM: I want to talk to you today about two key evolutions, let’s call it: The evolution of your work, the kind of questions that you’re asking, and then a big picture cultural evolution that’s been happening while you’ve been doing that work. And that’s pretty closely intertwined, you’re researching behavior, but I want to really talk about how your work is responsive to changes in society, in the big picture. How does that sound?
KM: Sure, I hope I can answer. We’re just doing our best, I don’t know! Not that responsive, probably, but we’re trying!
AM: Of course, of course. Okay, let’s start way, way, way back at the beginning. How did you become first interested in behavior change?
KM: Oh, gosh. It’s a good question. There are so many different ways I can answer that question. I think [at] first it was just as a person who’s introspecting, right? And I think that’s how everybody who’s interested in behavioral science or behavior change starts at the very beginning. It’s just like, “I’ve been noticing weird things that I do, that people I care about do”, and trying to make sense of the world. So that’s certainly the very beginning of it, is just being a human observer of the human condition, so you can call it me-search. But then I guess I would say more intensively, or the more academic answer, is that when I was an assistant professor here at Wharton, I wasn’t necessarily focused exclusively on behavior change, I would have said my identity was like, “I’m interested in decision making and all the weird things people do and all the mistakes they make, which is a pretty broad area.” And then, we have a med school here and I wandered over to med school, and ended up in a seminar room where someone was presenting this graph showing the proportion of premature deaths in the US that are due to different causes. So it was a nice pie chart breaking that down, it had things like accidents and environmental exposure and genetics, and it also had daily decisions on that graph. And the thing that really blew my mind and ended up being a pretty pivotal moment was seeing that that wedge about daily decisions was about 40% of premature deaths. It was bigger than any other contributor, and that really surprised me. And so I saw an opportunity to do something more impactful than I had appreciated with a focus I would have guessed like five percent…So the accumulation was bigger than I thought, and that helped me refocus and realize, “Okay, dabbling, studying decision making is one approach, but here’s a real opportunity for impact.” So that got me focused on behavior change, not just in health but also other consequential domains, like savings and education, where you can sort of imagine similar graphs. Maybe you haven’t literally seen them, but probably things accumulate more than I could appreciate it, which was the main takeaway I had from seeing that. So that’s what got me really hooked.
AM: Wow, 40% is a pretty staggering number. I don’t know if I would have pitched it at 40% just thinking about it.
KM: No, and I think it’s even higher today if you look at estimates because of the opioid crisis. And traffic fatalities have been going up as well because of texting and driving. That graph was from 2007, so I think the latest estimates are actually considerably higher, which is pretty extraordinary.
AM: Obviously, you’ve done a lot of really impactful work with the behavioral change for good initiative, and I was just wondering if you could give a little insight into the niche or the need you were looking to fill when you started that up with Dr. Duckworth.
KM: Yeah, sure. So that initiative actually was a response to a call for proposals that came out from the MacArthur Foundation, of all places, in 2016, basically saying “We’re gonna give $100 million to one team that can make meaningful progress on an important social goal. Tell us your best idea.” And there was an internal competition at the University of Pennsylvania and along with Angela – I like that you called her Dr. Duckworth – I was already really interested in behavior change, we were talking about it regularly, thinking about how we could work together, because we were both interested coming from different backgrounds and angles. And we were like, “Hey, what could be a more important problem than this, honestly? This is totally worth $100 million if you can come up with an innovative new approach to trying to make more meaningful inroads on these important problems”. So that’s really where it came from, was this external impetus, and we entered the competition. We actually did okay, I think we made it to the semi-finalist round. We did not ultimately get 100 million dollars. It went to Sesame Street to make programming for refugee children. And ultimately, actually, amazingly, they came to Angela and I like, “But we can’t figure out how to change their behavior!”, and we had a good laugh about that, but I love Sesame Street and I love Big Bird so no hard feelings. But it gave us our start and we got some seed funding from the University after they had sort of put us up as the one proposal through an internal competition. And the idea was sort of like, “Okay well, if we want to imagine we had $100 million to spend on making progress on this important issue, how would we optimally structure it?” Basically money is not a constraint under those conditions, what do we do? And we were like, “Well, let’s bring together all the brightest minds from different disciplines, have them collaborate, try to design, tournament style, different programs with different scientific insights built in and compete to see what works best”. Let’s partner with big organizations so we can run massive experiments and test things simultaneously and see what wins, and that’s really basically where the whole thing was born.
AM: What’s really interesting is what you just mentioned at the end there: running massive experiments at the same time. That leads into one of the biggest things that you guys have done at the behavioral change initiative, which is all your work with mega studies.
KM: That completely came out of this call, and then once we sort of dreamed that up, we were like, “Oh, we should definitely do this”. So we did it anyway, even without $100 million.
AM: Obviously now, it makes so much sense that money was a part of the origin for how you started this work, because your work with mega studies, it’s really done a lot to make behavioral change research more cost effective in the long term. And really, when I looked into your background with mega studies, what really fascinated me is how it really serves as a study about studies. It really changes the way we look at how we can conduct behavioral change research. Can you talk a little bit about why mega studies are so promising and when it makes behavioral change research more effective?
KM: Yeah, I mean we’re very excited about their potential. I also, by the way, think they have plenty of limits, and I’m glad that we still do work other ways too. The potential is, instead of throwing one thing at the wall and having it stick or not, we would like many things to be tested simultaneously, and we get comparable evaluation statistics about how impactful they are. We can also look at heterogeneity, so like what works best for whom, much more effectively when we test twenty things than when we test one at a time, and we can really start to hone in on, “Oh, for these subpopulations there’s real value in this approach. These other subpopulations really respond better to something else”. Coinciding with machine learning taking off, it’s nice to have data that allows you to look at those kinds of questions. Let’s see, I’ve loved the cross pollination, because normally everybody’s pretty siloed: economists publish in econ journals, psychologists [in] psych journals, marketing people in marketing journals, medical doctors [in] medical journals, and they don’t have as many reasons to rub elbows. In the mega study framework, each of those groups can still design their self-contained research study, but then they get pooled and glued together inside of a mega study, and they end up getting to see what other people are up to in different fields and co-authoring the ultimate megastudy paper, even though their own study gets to be its own research paper in their preferred journal. And so I think that’s been really nice too, just for breaking down some barriers, getting people to learn from each other who might otherwise not have known that there was a conversation to join. And the fixed costs are borne by a single organizer, so we can reduce the marginal costs for individual scientists with a cool idea who don’t have the wherewithal to set up a field site to test something policy relevant, so that’s been great too.
AM: Okay, skipping to the end here, I just want to ask you about Choiceology. How important is it that the everyday person knows about and understands behavioral economics?
KM: That’s a great question. How important, that’s a hard question to answer. I think there’s a lot of value in it, because so many of our decisions can be optimized or better if we understand decision biases. We can make better choices about our finances, our health, our education, our parenting, our job choice, our mortgages, if we get the science right. So the hope is just to bring a wider set of people the knowledge that means they’re not making mistakes that are costly. So it’s hard to quantify what that’s worth, but I think it’s worth a lot, and that’s why I do it.
AM: Thank you so much for your time, and I’d love to continue this conversation going forward.
KM: Thank you.