Probable Causation

View Original

Episode 50: Desmond Ang

Jennifer Doleac

Desmond Ang

Desmond Ang is an Assistant Professor of Public Policy at the Harvard Kennedy School.

Date: April 27, 2021

Bonus segment on Professor Ang’s career path and life as a researcher.

A transcript of this episode is available here.

See this content in the original post

Episode Details:

In this episode, we discuss Prof. Ang's work on how police violence affects local students' educational achievement:

"The Effects of Police Violence on Inner-City Students" by Desmond Ang.

OTHER RESEARCH WE DISCUSS IN THIS EPISODE:

TRANSCRIPT OF THIS EPISODE:

Jennifer [00:00:08] Hello and welcome probable causation, a show about law, economics and crime. I'm your host, Jennifer Doleac at Texas A&M University, where I'm an Economics Professor and the Director of the Justice Tech Lab. 

 

Jennifer [00:00:18] My guest this week is Desmond Ang. Desmond is an Assistant Professor of Public Policy at the Harvard Kennedy School. Desmond, welcome to the show. 

 

Desmond [00:00:26] Thank you so much. 

 

Jennifer [00:00:27] Today, we're going to talk about your research on how police violence affects the educational and psychological well-being of local students. But before we get into that, could you tell us about your research expertize and how you became interested in this topic? 

 

Desmond [00:00:40] Sure. So I'm an applied economist. My research focuses on topics of racial inequality, pretty generally, focusing particular on the intersection between race, government and education. And in terms of why I'm interested in this topic and how I tend to be interested in it - on a personal level, I grew up in the part of the United States where there just weren't very many people that looked like me. And so I've always thought a lot about and been interested in the idea and the extent to which our interactions with government, with institutions influence the sense in which we feel a part of society and the types of futures we still envision for ourselves. And I felt like that just wasn't something that was often reflected in the empirical literature or the academic literature, specifically in economics. You know, I think there's worth thinking about discrimination on the part of the discriminator, but less so thinking about whether the downstream ramifications for minority groups. And in thinking about that sort of general idea, I started to think about police killings because these are obviously very - extremely salient events. There's a long history of them in the United States. And there's just a ton of sort of anecdotal ethnographic works suggesting that these can be really traumatic for certain communities precisely because of this idea behind government and racial discrimination and perceptions of justice, etc. And so it seemed like a very natural setting to try to answer this question empirically. 

 

Jennifer [00:01:57] So your paper is titled "The Effects of Police Violence on Inner-City Students." It was recently published in the Quarterly Journal of Economics. And in this paper, you consider the effects of police violence in Los Angeles. So what type of violence are you focused on here and how prevalent is it in L.A.? 

 

Desmond [00:02:13] Yeah, that's a great question. So I focus specifically on officer involved killings. So instances in which generally an officer shoots a civilian, that person dies. It's rare that you know the person is killed through other means by the officer. These events are actually surprisingly prevalent. Probably more so than most people think across the United States. There's a thousand police killings a year in L.A., which experiences more than any other county. L.A. accounts for about four to five percent of that nationwide figure, so about 40 to 50 a year. So over my sample period, which spans from 2002 to 2016, there are over 600 killings in Los Angeles County. 

 

Jennifer [00:02:48] And so your going to be focused here on how those killings affect local students. So why are you interested in that outcome? Why might police violence affect local students? So in other words, what are the mechanisms here that you have in mind? 

 

Desmond [00:03:01] Yeah, I think there's a couple of main mechanisms. So one is that these events are violent and there might just be some dramatic effect of being exposed to the fact that somebody was killed nearby. So there's a lot of work sort of public health epidemiology looking at this correlation between adverse childhood events like you get in a car accident or you know someone that dies when you're really young and how that's correlated with later life outcomes. 

 

Desmond [00:03:22] Patrick Sharkey, a sociologist at Princeton, also has done really seminal work showing that sort of neighborhood violence is really bad for students' academic performance on high stakes tests and other outcomes. And so that's sort of like the violence angle. The other mechanism I have in mind is perhaps more anecdotal, something that we've seen recently with things like Black Lives Matter and the aftermath of killings of Michael Brown, George Floyd. Going back, you think about something like the Kerner report back in the 1960s and even earlier. There's been this concern among communities of color that the criminal justice system treats them unfairly right, and that these actions of the police specifically as sort of an arm of the government, a very visible arm of the government, they carry additional weight. It's not just about violence solely, but it's the sense that, you know, if the government is killing somebody and oftentimes there's not necessarily retribution for those things happening, especially if the victim is a person of color, then it might raise these concerns about racial discrimination, about procedural justice that again, we've come to a head recently where that, you could imagine being a person of color growing up in these communities that this might impact what you view for yourself, how you think society views you and then might have all these consequences, all these other domains that we haven't necessarily measured. 

 

Jennifer [00:04:33] So before this paper, what did we know about the effects of police violence? 

 

Desmond [00:04:36] We knew a lot, at least anecdotally. Right. So there's a lot of sociological studies, a lot of ethnographic work talking to young individuals of color growing up in these communities like - what's the experience of being stopped by police. Right. We also know from things like the Kerner Commission report that there's a lot of reason to believe that these events are just, like, really socially destructive and they heighten racial tensions. They heightened distrust in government. All these other spillovers from them. But in terms of sort of empirically, there was much less work quantifying these things. Right? So there's this work by some Harvard medical researchers looking at black mental health in states where police killings occur, they find sort of a negative correlation there. There's other work that focuses on some of these really high profile incidents of police violence. Matthew Desmond has a paper looking at police beating of Frank Jude in Milwaukee. He finds some evidence that 911 calls, particularly from black neighborhoods, fall after this event. There's work by Seth Gershenson and others looking at student achievement after the Michael Brown killing in the Ferguson riots. And there's a little bit of work in political science looking at the aftermath of the Rodney King riots and sort of the political aftermath of that. So we knew - we had those case studies and we had a lot of evidence to believe that this is something that communities of color were very much concerned about for a very long period of time. But we didn't have as much robust empirical evidence as you might expect, given what a big sort of topic it is for certain communities. 

 

Jennifer [00:05:59] So why don't we know more than we do? What are the hurdles that researchers like yourself have to overcome in order to measure the causal effects of these events on various outcomes that we care about? 

 

Desmond [00:06:10] Yeah. So that's a great question. In terms of why, you know, don't we know more? Why didn't we know more earlier? I think some of it is just sort of representation and who's doing the study and what questions we're interested in. But from sort of the more tactful standpoint in terms of identifying the causal effects of these events. Sort of the underlying concern is that, you know, police killings aren't random. You can imagine that they're more likely to occur in neighborhoods that are already disadvantaged. Neighborhoods that are already very heavily policed. You might also be concerned that you know they are occuring in the aftermath of some other event. So there's a hostage situation, and the police come in and kill that person. And all these other factors can be correlated with outcomes that we care about. Right. So students coming from disadvantaged neighborhoods are not only more likely to be exposed to police killing, but they also might have lower academic achievement to begin with. So that is sort of like the fundamental concern about, you know, how do you identify the causal impact of these events given that they're not random? 

 

Desmond [00:07:01] But it turns out that at least within the neighborhoods, once you really narrow in on the exact location and timing of where these events occur, that it happened on this street block, not the street over, that it happened on this week and not the week later, at that really narrow level, they are pretty close to random. At least that's what I find. And so that allows us to identify the causal effect of these events as long as we have really, really granular data so that we can look at the aftermath of these events in really short time horizons and very, very small geographies. Then you can ask questions like, well, how does attendance change in this block versus the block over in the days just before and the days just after where police killing occurred? And in that way, you can really, if you think about potential confounds and then - you'd really need to think of something that's happening, essentially at the exact same location exact same day for this not to be due to the police killing. And that becomes much less likely as again, you have this very granular information. 

 

Jennifer [00:07:57] And so you need this granular data. So what data are you using in this study? 

 

Desmond [00:08:03] So I combine the different datasets. The first is this really detailed data on the universe of Los Angeles Unified School District high school students from 2002 to 2016. And so getting to this idea of sort of geographic granularity, this includes each student's exact home address, so we know exactly where they live, which satisfy sort of this geographic component. And then obviously we have their outcomes so we can track their GPA, we can track whether or not they graduate from high school, whether or not they're diagnosed as certain symptoms of emotional trauma. And then for later, for some of these cohorts, we also have daily attendance. You know for a given student whether they attended school on sort of a given calendar date down to that level of temporal granularity. And we're going to combine this or I combine this with pretty detailed instant level information. So we know exactly for all these 600 police killings that occurred in Los Angeles over this 15 year time period, you know the exact date and location of the event, as well as other contextual information, which I draw from District Attorney Incident Reports. 

 

Jennifer [00:09:01] So just as an aside, this sounds like a difficult dataset to put together. Is there a back story here? How long did it take you to actually get all these data and convince people to give you this data? 

 

Desmond [00:09:11] I think I got - I think I got very lucky with this. So I applied for the academic data and through L.A., through their sort of standard protocols - and I was very fortunate in that they - there are people that were interested in the topic. Obviously, it takes them a while to gather all this information and hand it over. So that process probably took half a year. In terms of the district attorney stuff, that was an instance where for a lot of these events, I mean, there are 600 police killings, there's each of these reports is, you know, something like five pages long. And I was just reading them. You know, I was reading them and sort of hand coding, you know, was this person armed, were they not. What were the other factors going on? I mean, that took a while and it was also like very taxing work in a lot of different ways. But it was something that, like I learned a lot from. 

 

Jennifer [00:09:54] So what do these events look like? Who do they typically involve? Where do they occur? Are these events all, you know, reported in high profile media events? Or - just give us a sense of what events are in this data set? 

 

Desmond [00:10:05] Yeah. So these police killings, they disproportionately involve young minority men in terms of the person that was killed. So about a quarter of them - this is in Los Angeles, where people who were killed blacks, about 50 percent were Hispanic. This is overrepresented relative to those groups' population share more generally. So again, that, you know, on a nationwide basis, we also know that people of color, specifically males, are more likely to be killed than whites. 

 

Desmond [00:10:30] In terms of like the context of these events, about half of the people who are killed by police were armed with a gun. Put differently, that means about half of them either had no weapon at all or had some weapon that wasn't as lethal as a gun, for example, a knife or a pipe or something else. And what's kind of surprising is that most of these events, the vast majority of them weren't mentioned in media, at least not the media sources I looked at. So only 20 percent of them ever showed up in their print versions of six local newspapers, including the L.A. Times, which is, you know, a pretty big newspaper. Again, that could obviously be missing other forms of media like social media, online publications, etc. But it's really important to note just the disparity between something like this or the sort of the average killing I was looking at here and what you might see from these high profile killings, like the death of George Floyd, the killing of Breonna Taylor, where this is sort of just known on a national level. For most of these events, that's just really not the case. Despite the fact that, again, they often involve minorities and in a surprising share of cases, they involve someone that was also completely unarmed. 

 

Jennifer [00:11:30] So you're going to wind up using super local information about where the events occur and which students are living there to figure out who's affected by these killings. So in order to do this, you graph the effects on attendance for the students based on the distance from the incident and the time since the incident. So what do you find when you do this? 

 

Desmond [00:11:49] Yeah. So exactly as you said, you know, we have every student's home address. We know exactly where these incidents occurred. So you can calculate for each student how far they were from that incident, right, for the students that were enrolled at that time. And you can look at their trends in absenteeism in the week before this event occurred and the week after. And what you find is that in the week before, it's like a pretty constant relationship. So regardless of whether or not you live very close to the incident or you live further away like a few miles away from where the incident would occur, in the week before the incident occurs, like the rates of absenteeism are really similar across these groups. In the week after you see this pretty sharp spike in absenteeism for students that live specifically pretty close to the event. So within about half a mile of the incident, these students are significantly more likely to miss school in the week after the event. 

 

Jennifer [00:12:34] So as I said, you're then going to use that information on that precise timing and the precise locations of these police killings relative to where the students live to measure the causal effects of these killings. So walk us through your approach. What's the ideal experiment that a researcher would run to measure the causal effect of these events? And how does your approach approximate that experiment? 

 

Desmond [00:12:54] Yeah. I don't know if the ideal experiment that we'd want to run is the right one. But, you know, I think from a pure causality standpoint, you know, if you think about a randomized controlled trial like the idea is that you're going to randomly assign one group a treatment and one group no treatment, a placebo. Right. So you give one group, a dose of a new vaccine you're testing and on another group, you have something similar that's actually not a vaccine. In this situation, you kind of want to randomize who is exposed to these police killings and who's not exposed to these police killings. That's obviously not a thing that we actually want to do. We don't actually want to either, you know, actually have police go out and randomly kill people or even just sort of prime that in people because that itself could be sort of traumatic for individuals. 

 

Desmond [00:13:33] But that is sort of the general thing that we're trying to do and what we try to do - what we'll do instead is essentially take as random sort of within these neighborhoods within these really small geographies, the fact that it occured on this block versus another block and the exact timing of when it occured to sort of approximate this idea of sort of a random experiment. 

 

Jennifer [00:13:54] Yeah. It's a good flag that most researchers would not want to run this actual experiment. And even if someone proposed it, it would never get through the IRB. But it's always useful to think about what the random variation you might like to have. Okay, well, then let's talk about what you find. So what's the effect of exposure to a police killing on students' academic performance? 

 

Desmond [00:14:13] Yeah. So for students that live within half a mile of where these events occur, they experience a pretty significant decrease in their GPAs. You see their GPA drop .08 points on a 4.0 scale. And this effect lasts several semesters. So you'll see that there's a pretty big dip and it takes quite some time for students' GPA to recover after being exposed to one of these police killings. 

 

Jennifer [00:14:34] And then you're also able to look at a measure of psychological well-being. So first, tell us what that is and then what's the effect of exposure to a police killing on that outcome? 

 

Desmond [00:14:44] Yeah. So in the data, we know whether students are classified with something called emotional disturbance. And so this is a special education designation. It's sort of like a federally classified thing, and it's kind of hard to describe - it's correlated with PTSD, it's correlated to trauma. It's this idea that students have trouble learning or difficulty paying attention, and it can't be explained by, like an obvious health factor. So it's sort of like, again, this idea of like you've been kind of traumatized or you're having difficulty concentrating for sort of reasons unknown. But it is a pretty severe diagnosis because to be classified, there's - this costs the district some amount of money, because students with emotional disturbance have - get their own classrooms or they might even be set to different schools, etc., to accommodate these needs. And what I find is that being exposed to police killing those students are about 15 percent more likely to be diagnosed with emotional disturbance afterwards. And this lasts, I mean, this is a persistent diagnosis, so it lasts for quite some time. 

 

Jennifer [00:15:37] And to those effects depend on the race of the student? 

 

Desmond [00:15:39] So when we look at the GPA effects, they depend pretty heavily on the race of the student actually. So we can look for a given police killing like how does that affect the GPA of black students that live nearby, the GPA of the Hispanic students nearby and the GPA of like white or Asian students nearby? And what you'll see is the largest effects are coming from minority students. So Black and Hispanic students are significantly impacted by exposure to police killing whereas white and Asian students are not. And this again, this is not due to the fact that white and Asian students are less likely to be exposed right. We are conditioning on whether or not you live close to that event. So you were actually close to that event. It's just that these students respond very differently, even conditional on exposure. 

 

Jennifer [00:16:15] And do the effects depend on the circumstances of the police killing whether the person had a weapon, things like that? 

 

Desmond [00:16:21] They do. And so we also find that the strongest effects that come from police killings of unarmed individuals. And you kind of see this sort of linear relationship between, like how deadly that weapon is. So the moderate effects for police killings of individuals that were armed with something like a knife. And you see even smaller effects for individuals - police killings of individuals that have a gun. So it really does seem to suggest that the largest effects really are coming from police killings of unarmed individuals, who you can imagine might - those are the instances in which there wasn't some other crime that was going on. There wasn't some other violence going on that precipitated the police killing. 

 

Jennifer [00:16:56] And so you do those tests in the context of thinking through mechanisms. So how do you interpret those results - the fact that the effect varies so much with the race of the student and the circumstances of the killing itself? What do you take from those results? 

 

Desmond [00:17:09] Yes. I think this lines up with a lot of even recent events that we've seen, right. Where police killing George Floyd, for instance, and Black Lives Matter, right. The extent that certain communities are much more impacted by this in part due to these perceptions of racial discrimination and bias and historical discrimination in the criminal justice system and more generally. And that there's like - that this mechanism, in part, is driven by these sort of perceptions of injustice. So, you know, if we thought that all the effects were coming from the fact that there was just like a big shootout, right, that it's all just like gunfire, and that's what's driving student responses, then you would really expect to see the largest effects for police killings of someone who is armed with a weapon, right, where someone who is armed with weapon and shot at officers or civilians. And we really don't find large effects in those circumstances, but we do find really large effects in which police killed somebody who is completely unarmed. And the thinking there would be that, you know, if police hadn't killed this person, there's probably not going to be a shootout, there's probably wasn't a shootout that precipitated this. But what happens when these events occur is that you're activating all of these potential concerns about injustice, about discrimination, etc. And I think that very well feeds into and loads onto this aspect of race that we're seeing. 

 

Jennifer [00:18:17] Yeah. So you compare the effects of police killings with the effects of local criminal violence. So first, why do you do this? And second, what do you find? 

 

Desmond [00:18:26] So the reason why I do this again is to think about these mechanisms, the extent to which, you know, police killings are - they are sort of a bundle of things. It's the fact that police might have acted in the way that citizens might be concerned about. It's the fact that somebody was killed. It's the fact that there was gunfire, etc. And so to try to unbundle, you know, what is driving these effects? I look at the effects of some criminal homicides, right where somebody is killed, there's gunfire, etc., and how their students respond to that. And then that allows us to say, okay, well, if students are responding this way to gunfire or criminal homicide and in this way to the police killings, and there's some difference there, then that can tell us whether or not this is just violence or whether or not this is something to do with policing and injustice, etc. 

 

Desmond [00:19:07] And so consistent with, again, the work by Patrick Sharkey, work by others looking at the effects of neighborhood violence on students' outcomes like you do find that criminal homicides have a negative impact on students' GPA. It's significant. It lasts for some period of time. And in aggregate, you can imagine, you know, there's many more criminal homicides than there are police killings. In aggregate, you might imagine that those effects are larger. So the marginal level, like if you're only exposed to one of these things, which one has sort of the more damaging effect on student outcomes, you're going to see that the police killings have an effect that is about 50 percent larger than being exposed to a criminal homicide and that there is very much this racial component, which is to say that the race of the person who was killed by police matters a lot with regards to how students are responding to it, whereas the race of the person who was killed in sort of the criminal homicide doesn't really matter. You're seeing similar effects on student achievement, regardless of the race of the person who was killed in just like a gang shootout or something like that. And so in that regard, it again tells us something about the mechanisms, which is to say that it seems like for police killings, the mechanism is going to be perhaps very different from how students are responding to just local neighborhood violence. 

 

Jennifer [00:20:13] You also consider the effects of police killings on longer run outcomes, such as high school graduation and college attendance. What do you find there? 

 

Desmond [00:20:21] So I find that students who are exposed to police killings in high school are significantly less likely to graduate from high school or to enroll in college. These effects are about three percent of [00:20:29]the mean [0.1s] of students who were exposed to police things were about three percent less likely to graduate high school, two and a half percent less likely to enroll in college. And again, you're going to see similar patterns with regards to sort of the racial responses. So all of these effects are really loading onto black and Hispanic students, where they're much less likely to graduate from high school if they were exposed to a police killing whereas for white and Asian students, they're not any less likely or more likely to complete high school or enroll in college if they're exposed to one of these events. 

 

Jennifer [00:20:58] So that is your paper. Have any other papers related to the effects of police violence come out since you first started working on this project? 

 

Desmond [00:21:05] Yeah, there have been a few, which I think is great and very important. In other contexts,  there's a paper I read recently looking at a police killing in Chile and how that affects student protest behavior. In that setting, there's other work following up on - so the Matthew Desmond's paper looking at 911 calls after police violence, and they find actually results that are somewhat counter to Matthew Desmond's paper. And I also have a new working paper with John Tebes here at Harvard looking at police killings and how that affects civic engagement in Los Angeles. So, we're in the same setting using similar data, and what we find is that these events lead to increases in local registrations and voting among residents, particularly sort of the young minorities and Democrats. And then if we look at referenda data, we find that these events also increase support for criminal justice reforms aimed at sort of weakening penalties for minor crimes. So again, I think this is very consistent. All of this work is somewhat consistent with this idea that police killings are somewhat different from regular violence, right? And that it's activating some of these senses of injustice, concerns about criminal justice system, etc., and also ties in with recent conversations that we've had sort of publicly around the role of Black. Lives Matter and what that impact was for President Biden's victory last fall. 

 

Jennifer [00:22:17] It's really interesting that you're finding that police killings increase civic engagement, essentially because you could easily imagine it going the other way, right, that it's triggering concerns about legitimacy and whether your government cares about you. You could easily imagine that dissuading people from voting. You have any thoughts on why you're not seeing that? 

 

Desmond [00:22:38] Yeah. And I think you're totally right. And that was sort of my prior going in. Is that, you know, we're seeing students drop out of high school because of these things, and so you might imagine that people are so they're going to drop out from the electoral system or sort of civic engagement in that way. I think there's a few things going on. One is obviously it's a different sample, right? And that there's selection into who is a potential voter, etc. And that I think there's another aspect, which is that there's just heterogeneous effects, right. So you could imagine that there are people who are really skeptical of the government, who are unlikely to be registered anyways. These events occur, and that actually just increases their skepticism, right. Like they are now less likely to vote or less likely to register, even though they weren't registered to begin with. And so we're having sort of that marginal effect. And then there's these people who are sort of - especially who we've seen sort of partake in the Black Lives Matter movement, which are sort of like younger individuals who are still sort of coming of age, still grappling with these issues and these events occur and it sort of spur some of them to actually turn out more. So again, I think there's a lot that we don't know, you know, these are just average effects. Again, getting to this idea of heterogeneity and sort of the fact that people will respond differently and that we're only picking up sort of the average across all of that and really thinking about, you know, whether the consequences in some distributional way of these events on social cohesion, civic engagement, etc. that we just don't know. 

 

Jennifer [00:23:56] So what are the policy implications of your results and the other work in this area? What should policymakers and practitioners and advocates listening to this take away from all of these findings? 

 

Desmond [00:24:07] Yeah. You know, I think the first order thing is that the recent literature pretty clearly demonstrates is that, you know, police violence can have these negative social impacts, you know, in a way that might seem obvious in the ex post, but that we really have never really quantified or taken seriously despite a lot of concern raised by minority groups. Again, you know, historically empirical research has always focused on crime as outcome, right. So how does policing affect crime? Does having more police reduce crime? Does changing punishments reduce crime, etc.? And I think what we're saying here and what the literature is saying and what my paper is saying is that, you know, police play a really important and often very multifaceted role in communities, right. And the actions that they take that have spillovers on a whole bunch of other domains, not just crime, but things like health, education, voting, etc. And so I think the policy implication, rather than sort of the policy recommendation, is that these are all really important things that we should be weighing and quantifying because at the end of the day, you know, if you think about like, how should we be thinking about optimal levels or types of policing, we essentially want to weigh all these benefits, right. So we want to weigh, you know, pros on crime, some of these other domains, and really be able to aggregate that in a way that that makes sense. And so obviously, again, previously we focused on one side of the ledger. And I think this is really just taking us a step closer to getting this fuller accounting. 

 

Jennifer [00:25:27] Yeah. And what's the research frontier? What are the next big questions in this area that you and others will be thinking about in the years ahead? 

 

Desmond [00:25:34] I think there are a few things. You know, I think in a pretty narrow sense, if we're talking about law enforcement and policing, I think again, there's more work to be done for thinking about impacts on other domains, things that we might think of as being important when we think about, you know, how should we be allocating policing. So impacts on community well-being, impacts on income. And I think that there's also a sort of like broader discussions and work that needs to be done about thinking about like, how do you organize all this stuff, right. So at the end of the day, if what we want is we want to know the impacts on crime, we want to know the impacts of education and health and voting and how do we sum that up in a way that then gets to this policy implication? So I think that's one aspect of it again, in a pretty narrow sense, thinking about policing. Another aspect of it is thinking about like, how do we improve policing, right. And obviously, that's a loaded question, because again, we don't really know what the objective is. Like, how should we be measuring that objective? But I think that is sort of one direction in which the research will hopefully provide this guidance. And I think in a broader sense, you know, if my interpretation of these results is convincing, I think that we need to think more about these downstream consequences of just like these perceptions of discrimination, these perceptions of marginalization and discrimination, etc., on minority populations. So obviously, I'm looking at this very narrow sliver, which is these fatal police killings and the extent to which this might trigger some of those feelings. But this is like a vast thing, right. Like if you are a minority in the United States, you've experienced discrimination over some cumulative span of your life in a number of different domains. And how can we really think about what that cumulative effect is. How can we think about - how can we mitigate any negative consequences of those feelings, etc.? 

 

Jennifer [00:27:12] My guest today has been Desmond Ang from the Harvard Kennedy School. Desmond, thanks so much for talking with me. 

 

Desmond [00:27:17] Thank you so much, Jen. 

 

Jennifer [00:27:24] You can find links to all the research we discuss today on our website probablecausation.com. You can also subscribe to the show there or wherever you get your podcasts to make sure you don't miss a single episode. Big thanks to Emergent Ventures for supporting the show, and thanks also to our Patreon subscribers. This show is listener supported. So if you enjoy the podcast, then please consider contributing via Patreon. You can find a link on our website. Our sound engineer is John Keur with production assistance from Haley Grieshaber. Our music is by Werner and our logo was designed by Carrie Throckmorton. Thanks for listening, and I'll talk to you in two weeks.