The opioid crisis has not abated, but so far policy has remained far less punitive than for prior drug epidemics. Is that because it has been characterized and seen as a “white” drug problem? Could it take a more punitive turn if its connotations change? Tanika Raychaudhuri finds that sympathetic media coverage makes Whites more supportive of treatment over punitive policies for Black and White users, but less so for Black users. Justin de Benedicts-Kessner finds that White and Black Americans are more sympathetic when they see opioid users who look like themselves. Users who got into opioids through prescribed drugs see the least blame.

Guests: Tanika Raychaudhuri, University of Houston; Justin de Benedicts-Kessner, Harvard University

Studies: “The Political Effects of Opioid Addiction Frames” and “How the Identity of Substance Users Shapes Public Opinion on Opioid Policy


Matt Grossmann: How race makes us less punitive on opioid policy, this week on the Science of Politics for the Niskanen Center, I’m Matt Grossmann.

The opioid crisis has not abated, but so far policy has remained bipartisan in polarizing times and far less punitive than for prior drug epidemics. Is that because it has been characterized and seen as a white drug problem? Could it take a more punitive turn if its connotations change?

This week I talk to Tanika Raychaudhuri of the University of Houston about her Journal of Politics article with Tali Mendelberg and Anne McDonough, The Political Effects of Opioid Addiction Frames. She finds that sympathetic media coverage makes whites more supportive of treatment over punitive policies for Black and white users, but less so for Black users, they face less sympathy.

I also talked to Justin de Benedictis-Kessner of Harvard University about his political behavior article with Michael Hankinson, How the Identity of Substance Users Shapes Public Opinion on Opioid Policy. He finds that white and Black Americans are more sympathetic when they see opioid users who look like them. Those who got into opioids through prescribed drugs, see the least blame from everyone.

We focus on the public response today, but both review the now substantial evidence that media coverage of the opioid crisis has been quite racialized. Not necessarily matching the facts on the ground, but perhaps opening the space for more effective opioid policy. Let’s start with Raychaudhuri, who experimented with common media narratives to see their effects.

So what were the major findings and takeaways from your article about opioid addiction frames?

Tanika Raychaudhuri: So our paper explores whether the ways that the media frames the opioid crisis shapes public opinion about the appropriate policy response. So prior research shows that the media frames opioid use sympathetically as a public health issue that primarily impacts white Americans. In our paper, we explore whether the sympathetic valence or racial frame of media coverage of opioid use impacts support for public health or punitive responses among white Americans using randomized survey experiments. And our key takeaway is that the sympathetic media frames of the opioid crisis have increased support for treatment, but there are some differences in the level of support for treatment based on the racial group featured. So sympathetic frames of white users, particularly increased support for treatment relative to those that feature Black users.

Matt Grossmann: So let’s back up a little bit for those who may not be familiar and talk about just what is the state of the opioid addiction problem in the US? Is it still growing? What are the real racial patterns of use and abuse? And maybe just what we are talking about when we talk about punitive versus other kinds of responses.

Tanika Raychaudhuri: Absolutely. So I think the opioid epidemic is a very pressing public health crisis today. According to the CDC, over 645,000 people have died of opioid overdose deaths since 1999. And these death rates have increased, particularly in recent years, especially since the onset of the COVID-19 pandemic. And on average, I think the most recent statistic from 2021 is that about 220 Americans die per day of an opioid overdose.

When it comes to the racial patterns of use, white Americans are disproportionately impacted, but opioid use and overdose death rates are also growing rapidly among various communities of color, including African-Americans, Native Americans and Latinos. So to give one specific example, a 2022 Pew Research study finds that the sharpest increase in overdose deaths between 2015 and 2020 was actually among Black men. And in fact, as of 2020, the overdose death rate per a hundred thousand people in the population was actually higher among Black and Native American men than among white men. So the racial demographics are definitely starting to shift in the sense that opioid use is a problem that’s also impacting racial minorities.

So in terms of policies, I think punitive policies related to drug use historically have just been related to incarceration. So things like mandatory minimum sentences for drug possession, jailing people for using drugs and for having drugs on them, things like that. In sharp contrast, treatment oriented policies are really focused on helping people to find the tools to deal with addiction and to become less reliant on opioids. So some examples include the fact that there are opioid treatment centers that are often funded by the states where people can go to seek treatment when they have severe addiction issues. There’s also several pharmaceutical drugs available these days, which are in some ways like synthetic forms of an opioid or some kind of medication basically to help people to wean off of the drug and become less reliant. So I think the policy state these days is largely focused on making those types of drugs more accessible to ordinary patients and helping doctors to make sure that they have the tools and resources to be able to prescribe them.

Matt Grossmann: So you said previous research has found that the media racialized this drug problem as more specific to white Americans. So describe I guess what that research finds, and maybe if it has any kind of trajectory. Is it becoming less seen as a white drug problem in media coverage more recently?

Tanika Raychaudhuri: Yeah, so there had been previous research, particularly content analyses of news coverage, which shows that historically stories about opioid use disproportionately feature white opioid users. So the types of studies that focus on this topic often look at newspaper coverage, coverage in other media sources and measure the racial, I guess, demographics of the people who are featured in those articles or the groups that are described as being impacted. I haven’t actually seen any recent studies which suggest that there’s been a shift in coverage, but just anecdotally looking at the media these days, it does seem like there has been coverage of the fact that various minority communities are being impacted by opioid use. So I do think that in the future I would not be surprised at all if we see studies that point to that shift.

Matt Grossmann: So tell us a little bit more about the setup of your experiment. What were the actual treatments that people were exposed to and how did they compare to the actual news coverage that people might experience?

Tanika Raychaudhuri: So in our experiment, we tried to approximate actual media coverage of opioid use, and the primary experimental manipulation that we used in the study was a vivid and detailed hypothetical news story about opioid use. So rather than using simple vignettes, we tried to draw on real news coverage for wording and imagery. And the news stories that we used began with some statistics about opioid use and the growing opioid crisis and ended with the personal story of a specific person who was using opioids named Mike. We created four versions of that hypothetical news story as our experimental treatment, and we varied two dimensions of the story. So we varied what we call the valence of coverage, which is whether the opioid use was framed in a sympathetic light or an unsympathetic light. And we also varied the race of the opioid users featured in the article. So whether they were primarily white Americans or African-Americans.

And then the way that the experiment worked is that respondents were randomly assigned to view one of these four treatments. So we had a sympathetic version that featured white users, a sympathetic version that featured Black users, an unsympathetic version, which featured whites and an unsympathetic version that featured Blacks. And respondents were randomly assigned to view one of those four treatments or a control condition where they did not view any news story. And what we did is we manipulated language and imagery across those hypothetical news stories, but just to vary those two dimensions, so the valence of the coverage and the race of those who were featured, but otherwise the stories were identical across the experimental conditions. And one of the reasons that we use this approach rather than maybe a short or simple vignette is we felt that this would be more realistic and more likely to approximate real news coverage that people might come across when reading newspapers or magazines in the real world.

Matt Grossmann: So you find some big differences by how sympathetic the portrayal is, and then differences also based on race. So how do you interpret the relative importance of those findings and how should we fit them together?

Tanika Raychaudhuri: Yeah, so I think that’s a great question. And so what we look at in the paper is the impact of viewing these hypothetical news stories on support for treatment over incarceration, support for candidates who promote treatment and support for or willingness to pay increased taxes for treatment. And what we find in general is that these sympathetic frames are really strong and increase support for treatment and for candidates who promote treatment over incarceration.

So in general, these sympathetic treatments increase support for both of these political outcomes regardless of whether the story featured a white or Black user. So relative to not viewing any story, viewing the white sympathetic story, increased support for treatment by about 21 percentage points and viewing the Black sympathetic story increased support for treatment by about 13 percentage points. So we take this as saying generally the sympathetic frames of opioid use are effective and increased support for treatment.

However, there is a difference in the strength of that kind of sympathetic treatment effect in stories that feature white and Black users. So sympathetic frames that feature white users have a stronger effect on support for treatment than those that feature Black users. And we take this as evidence of what we call racially selective sympathy. So it’s this idea that generally sympathetic frames work and I think that’s kind of the most important takeaway, but there may be differences based on the race of users portrayed. And I think it’s important to note that our study really just focuses on public opinion among white Americans, and this finding may not hold if we look at other racial groups.

Matt Grossmann: And you’re also able to explain some of the differences between viewing Black and white opioid users on the basis of the racial resentment questions rather than partisanship or ideology usually. So tell us about those findings.

Tanika Raychaudhuri: Right. So when we were looking at what explains these effects, we were interested in exploring whether people’s racial predispositions or their partisanship and ideology might condition those racially selective sympathy effects. And what we find is that racial resentment and holding negative stereotypes of African- Americans condition those racial sympathy effects. So those who are high in racial resentment and hold negative stereotypes of African-Americans were more likely to respond negatively to the sympathetic frame of Black users relative to that sympathetic frame of white users.

And in contrast, when we conducted the same tests looking at whether there are differences between Democrats and Republicans or liberals and conservatives, we didn’t find any evidence that those factors condition the effects. So we take this as evidence that racial attitudes rather than political views are driving those racially selective sympathy effects.

Matt Grossmann: So many survey experiments have been found to find larger effects than we find when we try to actually implement a public education campaign or something in the real world. On the other hand, people might have experienced this same set of frames over years and repeatedly in this issue. So how should we think about whether these kinds of findings would apply to real world engagement with media coverage around this issue?

Tanika Raychaudhuri: Yeah, so I think that that’s a really important question for us to think about as researchers when we’re conducting these studies that may have implications for politics in the real world. I think that in general, people are affected by media frames in the real world because media is often the primary lens through which people understand political issues, especially when it comes to issues that they don’t have personal experience with. But that being said, I think that people’s reactions to media frames of issues like the opioid crisis are the product of the culmination of a long-term process of exposure to many different news stories that takes place over the course of months and years rather than a single day. So in turn, viewing a single news story about something like the opioid crisis in the context of this one shot experiment is unlikely to fundamentally alter people’s existing views.

And for example, just to kind of speak to that limitation, I think of this kind of study design, in our study, we actually had a fairly hard time framing opioid use in an unsympathetic light, so kind of framing it as a moral failing of opioid users rather than a chronic and systemic problem that is caused by the drug industry and doctors pushing prescriptions. And this was despite the fact that we did try to use fairly strong negative language in our treatment, and we took this as evidence that a single unsympathetic story is not going to be able to counteract exposure to many years of sympathetic coverage of something like the opioid crisis. So I do think that media frames matter, but I think the ways that we try to study them, especially using these types of one-shot experiments, may not capture that full effect.

Matt Grossmann: So on the one hand, this replicates a longstanding understanding of American drug policy and opinion, going back to the crack cocaine versus powder cocaine distinction in the 1980s and how those were racialized. On the other hand, there’s been a fairly dramatic liberalization of public opinion on drug policy since then, especially in support for legal marijuana, but also surrounding treatment policies in general. So how should we think about this in relation to the history? Is this to show that not much has changed or is this a different world?

Tanika Raychaudhuri: Yeah, so I think that’s a really interesting point. And as you said, public attitudes towards drug policy have changed a lot over the years and have generally liberalized since the 1980s and nineties when drug use was framed in this stigmatizing way by the media and political elites, and most of the policy initiatives to counter it were really focused on criminalization and incarceration. One thing I think is interesting though, is when we compare attitudes about marijuana use and opioid use, is that these issues are viewed very differently. So I think policy views on marijuana have shifted towards supporting decriminalization and particularly supporting legalization for recreational and medicinal use. Whereas on the other hand, the liberalization of attitudes about opioid use is more related to acknowledging that this is a public health issue that should be diagnosed and treated. So while perhaps there may be two sides of the same coin in the sense that people’s attitudes are changing fundamentally on drug policy, these are different issues that are associated with very different policy responses.

Matt Grossmann: So in the debate over the crack versus powder cocaine issue, it was a nice clean break where we’re literally talking about the same drug that has different racial communities associated with it. But this is, as you said, a pretty different situation in terms of how the drug epidemic came about. One way to interpret your findings is that this is part of evidence that opioids are really being treated differently because of their racial connotations, but as you said, a lot of other variables might make our treatment of that different.

Tanika Raychaudhuri: Yeah, I think that that’s definitely the case. I think that the racial framing certainly could play a part in things. And as you had mentioned, the drug epidemic of the 1980s and nineties related to crack and powder cocaine had that clean-cut racial distinction. With opioid use, initially, maybe we saw that this was primarily a problem that was associated with white Americans, especially when opioid use was particularly related to prescription drugs. Now we’re seeing a shift where a lot of the issues related to the opioid crisis are related to synthetic opioid use and racial minorities are being impacted. So perhaps we will start to see a shift where support for treatment will decline, but I think it sort of remains to be seen as this is still an emergent issue.

Matt Grossmann: So the other article for this episode also finds racial differences based on the identity of the users in support for treatment versus punitive policies, but it is able to show across racial groups that people favor their own racial group, and it has a comparison that it’s not about geography or gender. It is about in- group favoritism. So how would you relate those findings to your own?

Tanika Raychaudhuri: Yeah, so I really enjoyed reading Justin de Benedictis-Kessner and Michael Hankinson’s paper on this topic. And I think that the findings are in many ways complementary with the findings from our paper. So I think our findings relate to theirs in the sense that we find that white Americans, the primary group that we focus on in the experiments, are especially supportive of treatment when opioid use is framed as an issue that disproportionately affects members of their own racial group. So one interesting implication, drawing from their research, might be that this might be an effect driven by in-group identification rather than outgroup animus. And perhaps if we were to look at other racial groups like African Americans, Latinos, and perhaps Asian Americans, we might see a similar effect when they’re looking at their in- group rather than just a straightforward effect like this if we looked at the population as a whole and focused our coverage on white Americans.

Matt Grossmann: They also have a piece where they talk about how someone got into the opioid issue. Any sense that that might also be an important feature, that basically if this is viewed as people were tricked into it by their doctors or that this was the fault of a medical professional, that that’s the key issue?

Tanika Raychaudhuri: So I think our study implies that that is a really important dimension. Our sympathetic frame of opioid use really described the trajectory of one particular opioid user as starting when he was prescribed opioids as medication for a long-term back injury, and that he became addicted. He had sought treatment multiple times, but was unable to receive it. I think this reflects the realities of many people who suffer from opioid use disorder. For example, I think I’ve seen some data that less than 25% of people who have dealt with these issues have effectively been able to receive treatment, so I do think that that framing really matters and shapes the way that people view this issue.

Matt Grossmann: Is there a shift in opioid policy toward a more punitive approach now? And is public opinion playing a role in that, or is it polarizing along partisan lines? What are we seeing in terms of the real policy world?

Tanika Raychaudhuri: Yeah, so I think that that is a very interesting question that’s difficult to answer. So on the one hand, I do think that public support for treatment is still fairly high when looking at recent public opinion polls. And at the elite level, the opioid crisis is one of the few political issues on which we’re seeing some bipartisan legislative efforts, both at the state and federal levels. So for example, as recently as this past December, Biden signed an appropriations bill, which included a bipartisan Mainstreaming Addiction Treatment Act, which had some initiatives in it to make treatment more accessible to ordinary Americans. But there is some evidence on the other hand that the public may be becoming a little less concerned about the opioid crisis than they were at least a few years ago.

So to offer another example, there’s an interesting analysis from the Pew Research Center from 2022, which indicates that Americans’ concerns about drug addictions are declining even as opioid overdose deaths increase. So while this is speculative, it could be possible that those changes in public opinion are fueled by changes in the racial demographics of who’s being impacted by this issue, and perhaps we will start to see a shift in terms of the types of policies that are being proposed to a more punitive approach. When it comes to the potential that this is becoming a more partisan issue, I think it still remains to be seen how that’ll play out, but I do think that the opioid crisis is a distinctive issue in the sense that it is a little bit less polarized along party lines than many other political issues today, including other health issues like the COVID-19 pandemic, which became incredibly polarized when that was unfolding a few years ago.

Matt Grossmann: So how do you think about the, I guess, moral issue here of, one way of putting it is we can get bipartisan and popular policy if we basically use the public’s racism and their desire to only help sympathetic figures? That would suggest from a political professional perspective that they should use those attitudes in the public to gain widespread support. What do you think?

Tanika Raychaudhuri: Yeah, that’s definitely an interesting question. And I do think that when it comes to it, this research suggests that these sympathetic frames of users, whether they are from the majority racial group or from a minority group, are very effective at increasing support for treatment. So I think that, on a more positive note, maybe that would be a key takeaway. But I do think that this implication that when these frames feature members of the racial majority group, that they elicit the most sympathy, at least among white Americans, also may play an important role in that. However, I do think that as the country diversifies and the electorate diversifies, it’s important to note that these types of patterns may not hold for racial minorities. So it may be the case that we’re seeing this strong trend among white Americans, but things may look very different when we’re looking at African Americans, Latinos, Asian Americans, and other minority communities.

Matt Grossmann:[inaudible 00:23:33] study focused on white Americans’ attitudes. The Benedictis-Kessner looked across the racial divide, but also studied the effect of media narratives, finding remarkably consistent results.

So tell us about the major findings from your article on opioid user identity and opinion.

Justin de Benedicts-Kessner: So the main takeaway from our article is people are more likely to favor treatment-based policies, so policies that deal with opioid use disorder or OUD in a sympathetic way that focuses on medical treatment, when they see media stories about people from the same racial group as them. And what this basically means is people are more sympathetic towards people of their own racial group, and they’re less likely to favor a punitive policy that would involve law enforcement. We see some effects that are moderated by how blame-worthy the individual who’s profiled in these media stories is, some of which has to do with the fact that people perceive those who started using opioids via prescription pills, Oxycontin, as more sympathetic than people who started by using heroin.

Matt Grossmann: So let’s back up a little bit and tell people about opioids overall. How big is this problem? Is it still growing? And are there actual racial patterns of use and abuse of note, or is this just a media storyline that this drug problem became racialized differently?

Justin de Benedicts-Kessner: Yeah, so the opioid crisis is still growing, unfortunately. It’s not going anywhere. I think a lot of people thought that during the pandemic it would decrease in its size, people wouldn’t be overdosing as much because there was a worldwide pandemic happening. That isn’t true. So I think the data from 2020 show that this is still the leading cause of death for Americans under age 45, though I should probably get those figures updated. And a lot of the public health literature on this has shown this is just not going away anytime soon, especially given the rise in synthetic opioids rather than other types of opioids. Synthetic opioids like fentanyl, which you may have heard about in the news, are especially potent and therefore especially likely to lead to overdoses. So this isn’t a new phenomenon, but I’d say the current opioid crisis has really permeated the media with stories about substance use addiction. And this is highlighted for a lot of people the fact that they just didn’t realize this was happening to all types of people.

And I’d say the objective racial patterns of it are, we don’t have great data on use of opioids by race, but we do have data on opioid overdose rates normalized by the population. So compared to say every 100,000 people, how many overdoses are happening among Black Americans, among white Americans. And the rates of opioid overdose are much higher among Black Americans, which I think that most people don’t realize, in part because of the media stories that have surrounded this opioid crisis. It’s been predominantly accompanied by stories showing that white people are affected by this crisis and that especially people in really, I’d say, non-traditional locations for drug crises are also affected by this, so people living in more rural areas like West Virginia or the Rust Belt more generally.

And I think that’s because a lot of the compelling stories that first people were hearing about in terms of opioid use came out of these areas. There are a lot of great books by journalists and others about what’s led to some of the opioid crisis in areas like, say, West Virginia. But given the now really widespread availability of things like fentanyl as well as heroin, it’s not just confined to those areas, even though stuff like oxy and prescription oxy may have originally been much higher in those types of locations. So it became racialized as this white or rural drug problem I’d say mainly through the news media. It’s not in line with how the data on the ground has played out over time since this crisis really started becoming more serious, which has been going on for now many years at this point. Mike and I started working on the opioid crisis and policy surrounding it back in 2017 when it was just starting to get talked about, I think, in a lot of policymakers views, especially at the national scene, and it hasn’t stopped being important, and it wasn’t really that early when Mike and I first started working on these projects.

I should mention one of the really cool things that we aren’t able to do in article, but we were building off in our article, is just this objective fact that the media portrayal of the crisis has been super racialized. And we know that from other really good work by people like Brielle Harbin who’ve tracked these media stories over time, which is just staggering if you look at it. She’s done a bunch of media content analysis showing how different from reality in terms of the racial overdose rates the media stories are, is kind of wild.

Matt Grossmann: So tell us about the setup of your experiment here, what were people actually exposed to? I know they saw pictures as well as texts and how you can sort of be sure that they saw the treatment.

Justin de Benedicts-Kessner: The way this setup worked in our survey experiment was that we created these articles that highlighted a story of a substance user alongside a picture of that substance user holding some drug paraphernalia. As I said, Mike and I first became interested in the politics of the opioid crisis about seven years ago, each of us for different reasons. I moved to an area of Boston in 2017 that’s really still the center of the visible effects of the opioid crisis. It’s right next to a road called Massachusetts Avenue. Many people have referred to in the news as well as in politics as Methadone Mile, or perhaps more sympathetically recovery row, because of its concentration of treatment clinics, a food bank, a safety net hospital, several other social services. So that was how I became interested in the opioid crisis.

But for most people in the American public, I think hearing about the opioid crisis, the way they’ve heard about this has been, like for many other social problems, via the news media. So we wanted to test how these portrayals that appear really quite commonly in the media have been affecting the politics of the crisis, specifically public opinion about this issue. So we created these articles by, both of us have read a lot of media coverage of the opioid crisis, and one thing that we noticed was a lot of the stories that appear in the news media about opioid use and overdose rates are not just reporting on the statistics behind this. They’re reporting individual stories of people affected by this, people who’ve experienced overdose, people who have opioid use disorder, what the effects have been like for their family members.

And so we wanted to create articles that really were quite like the reality of the news media’s portrayal of the crisis. We did this focusing on individuals, we also did this using language specifically like those that was used in news articles about this. I started doing experiments when I was a graduate student working with Adam Berinsky and learned really well from him, the way to create an experimental treatment is just figure out exactly what’s happening in reality. If you want to have any external validity with or to make our treatments as generalizable as possible, you should really mimic what people have in the real world.

So we created these articles to look as realistic as possible. They included language about where the person was located. They included language about what they first started using opioids, what type of opioids they first started using, so whether they were using prescription oxy or illegally obtained oxy or heroin. And we also manipulated things like their gender using names as well as the pictures, to try to see how these different attributes of their identity affected people’s opinions. We were pretty sure that these manipulations worked. We asked what’s called manipulation checks, we basically just asked people what their perceptions of these various attributes that we were manipulating were. So what race do you think the person portrayed in the article was after they’ve read this article? What gender were they? Where were they located? And we saw pretty good recall rates of all these factors, especially for race. So that gave us confidence in the fact that these articles are actually manipulating what we think we’re doing.

Matt Grossmann: So you find different effects for white and Black respondents, and it might be unique to your article that you’ve had this Black respondent finding that Black respondents are more sympathetic to the Black story of the Black user. So talk a little bit about how you think about those findings and how they, I guess, differ from how it’s usually talked about as a white respondent, white racialized issue.

Justin de Benedicts-Kessner: Most of the stories of the crisis, and I think some of the mainstream media’s interpretation of the racialization of the crisis, has been about how white people are more sympathetic towards opioid users than say past drug epidemics. But we wanted to see the American public as a whole, how do they feel about the opioid epidemic? And so we got a nationally representative sample that’s not just white respondents, but also includes Black respondents. And I think if we were doing this again, we probably would’ve done a Black over sample, just to try to get a little bit less noisiness or variability in our estimates among the Black group of respondents.

But what we see is this strong effect among Black respondents in favor of treatment focused policy and not as in favor of enforcement or punitive policy when they see a Black user profiled in this media story. We see similar effects among white respondents, they’re less supportive of treatment spending when they see a Black user rather than a white user. And there’s not really many differences in terms of their support for the punitive policy based on race of the individual profile in the article.

And we take this as really good evidence that among Black respondents, they’re going to respond pretty strongly to enforcement or punitive policy that could further the racialization of the criminal justice system, and especially the way that Black Americans have been targeted by enforcement over previous drug crises. We know that the in-prison population, people who are incarcerated, there’s really big racial differences in who’s arrested and who is then incarcerated as a result of drug use, for not just opioids. And so perhaps the reason we are seeing these bigger effects among Black respondents is just that they are more aware of this racialization of the criminal justice system. And so especially for a more punitive policy, they’re not going to want to see Black individuals being targeted for those types of policies for opioid use.

Matt Grossmann: So you also did these other manipulations of gender and geographic residents and didn’t find much of an effect there. Were you expecting to find something in terms of the geographic effects? And I guess how should we interpret this as it’s racialized but not really about rural versus urban?

Justin de Benedicts-Kessner: Yeah, we really expected to see that the location of the people who were profiled in these media stories would matter, given how much these media stories in reality have focused on much more rural populations. I think people perceive the crisis, the opioid crisis in reality to be just happening in rural places, but it’s also happening in cities. And so we can talk about that more, but it’s kind of interesting how the media’s portrayal of this crisis has not matched reality. So we really expected that people will be more sympathetic towards these rural populations. But I think that part of the reason we thought this and many other people might think this, is just that rural identity and race are really conflated in a lot of especially white Americans views.

So the main effect, what we would say where rural versus urban locations might affect people’s views on policy is coming through race. So while people who live in rural locations are more sympathetic or more willing to support a treatment focused policy to confront the opioid crisis when they see people who live in similar locations to them, this might be just because we’re also varying race at the same time. So if we just varied their location, that might affect people’s treatment policy opinions. But that’s probably because when you say someone’s from a rural location, people assume that they’re white. When you say someone’s from an urban location, people assume that they’re Black, or more likely to assume that.

And so by netting out the effective race, we’re able to show that it’s maybe not coming as much through these stories of just rural individuals, it’s coming through these stories of white rural individuals that people are becoming more supportive of the crisis.

Matt Grossmann: So you also find pretty large differences based on whether the person is perceived as or portrayed as starting from prescription drugs versus especially heroin, but even starting from OxyContin on the black market. It seems like that might actually be a bigger takeaway for professionals or has been, that they’ve really been trying to portray this as an outcome of prescription drugs, because that’s sort of a key factor in how the public thinks about this issue. How do you compare that with the racial connotations that you’ve been talking about?

Justin de Benedicts-Kessner: Yeah, you’re right. People are much more likely to see a portrayal of an individual in the media as starting to use opioids via prescribed oxy. That isn’t necessarily the truth. A lot of people are not getting prescribed oxy, but it’s a really effective villain in this story. That’s why I think we’ve seen many of the policy makers focusing on say, the legal cases against pharmaceutical companies and the people distributing opioids via prescriptions, because of the way they’ve sort of misused the medical system. But in reality, a lot of people are not getting exposed to opioids initially through these prescriptions, especially in locations that are not the rural heartland that is talked about in a lot of the journalistic media portrayals of the crisis.

So we wanted to test this, like is actually affecting how people are supportive of policy, but specifically, how does this affect the way that people are blamed for what they’re doing? If they’re prescribed a pill, they become addicted because it’s an addictive substance, and it’s chemically addictive. Maybe people don’t see them as blameworthy as if they had started using drugs illegally to start with. And so we tested this and what we see is exactly what I’m talking about here. People perceive individuals as much less blame worthy when they’re legally prescribed oxy and then start using other drugs illegally. Between 5 and 10 percentage points less blame worthy than those substance users were described as first using heroin.

And so we take this as a pretty important piece of how people have formed perceptions about the opioid crisis and probably have formed them over time. So people coming into our experiment are already being treated by many news articles in the real world, and so they probably are starting with a pretty high level of sympathy and less levels of blame for individuals who are affected by the opioid crisis. Essentially, we think that’s important, but not the only part of the story here.

Matt Grossmann: So many times in survey experiments, we get findings from asking someone directly about something after we’ve shown it to them, but it turns out to not necessarily translate into kind of say, an effective public health intervention that people actually see in the real world. On the other hand, these are treatments that people might’ve been, as you said, repeatedly exposed to over time. So how do you think about the validity of this as really being part of the explanation for how people see the opioid crisis?

Justin de Benedicts-Kessner: So not all people will, in the real world, read depictions of individuals suffering from substance use disorder in the media. I think the effects we observe are going to have different effects when consumed in a real world setting. Some people are going to opt out, for instance, of reading these types of stories of people experiencing opioid use disorder or people who’ve overdosed. Not everyone’s going to read those.

That said, most people who do read them are probably going to read multiple stories. These are not stories that are alone. I talked earlier about my neighborhood in Boston being called Methadone Mile. There’s a whole issue of the Boston Globe that was dedicated to the opioid crisis with this big front page image of someone on Methadone Mile with this big headline Methadone Mile. So most people who read about the opioid crisis are not just consuming one story. They’re consuming multiple, they’re probably following along with it. And there are a lot of great journalists who are doing work to portray these types of individuals over time and continuing to do so.

So I’d say while we observe these effects of especially racial identity on people’s policy opinions, we’re really coming into this with people already experiencing a lot of media coverage of the crisis. We’re nudging people in one direction via one story in this one shot sort of article and treatment, and then immediately measuring their opinions. But I think there’s a much broader potential for public opinion changes due to exposure to media stories when they’re reading multiple ones. And especially over time, if you’re getting repeated treatment of any type of experimental treatment that’s a news media story, I think you’re going to see a much more durable effect rather than something that just dissipates over time.

Matt Grossmann: But the other context of the change is just the big liberalization of attitudes. I mean, drug policy is kind of on par with gay rights in degree of change, public opinion overall. So in some ways the opioid crisis was a cause, but in some cases it could have been an effect of the fact that we’re just much more sympathetic than we used to be to drug use and much more negative about the punishment approach. The war on drugs. So-

Justin de Benedicts-Kessner: It’s a chicken and egg problem.

Matt Grossmann: Yeah. To what extent did opioids just come at a time in which both public opinion and media coverage were already turning in a more sympathetic and less punitive direction?

Justin de Benedicts-Kessner: Yeah, it’s hard to disentangle the two for sure. When people like Chris Christie are in support of much more treatment-based policies for all types of drug use, you know you’re doing better when policymakers across the political spectrum are doing that and admitting that the war on drugs did not work. I think that’s a success story in general. I think there’s still a lot of room for debate, and we see this in how people have pursued different policy paths to confronting the opioid crisis. So some states have really focused on the legal case to try to pin this down on the manufacturers of Oxycontin as well as some of the people who distributed pills, some of the doctors who were prescribing them at rates that were not proportional to the population at all. But some states have much more focused on how are we providing services to these populations that are affected on this.

So even within a… I’d say a policy sphere where we’re much more united around less punitive policy now than we were 30 years ago or 40 years ago, there’s still a lot of room for disagreement. And so I think that what we are getting at in our article is, is this being nudged at all by some of the racialization of the opioid crisis. It’s not the first crisis that’s been racialized. So like marijuana use, I think people are much less perceiving this as a black drug use problem than they are perceiving it as pretty much across the spectrum. People are using marijuana illegally or legally. And so these things accumulate over time. People start to see, oh, okay, drug use is not just a thing that affects people in urban locations who look different from me.

Matt Grossmann: So the other article for this episode also looks at a racialization of this crisis, but with just interviewing white Americans, but they emphasize that sympathetic portrayals can work for everyone. They just don’t work as well for a black user. How would you relate those findings to Europe?

Justin de Benedicts-Kessner: So people come into reading an article whether it has a sympathetic or not as sympathetic of a frame like those they use in that article. They come into those with their existing racial biases. And so the main way I’d say our articles speak to each other is just racial attitudes probably put a ceiling effect on the impact of sympathetic frames. So you’re only going to be so effective portraying a substance user as really sympathetically as possible given that people come into this with some racial biases in how they portray these users.

Matt Grossmann: So you also have a previous article about the role of self-interest in the opioid crisis. And in particular, you find, I think people kind of unwilling to actually have opioid clinics near them. So that might suggest that the sympathy only goes so far. So how would you relate what you find here to your previous work?

Justin de Benedicts-Kessner: Yeah, so this previous paper that Mike Hankinson and I wrote really was motivated by some of the dynamics that are happening in many American cities. So taking away this sort of rural versus urban dimension of it. Within cities, there are these debates about where to place treatment clinics, and that’s really about implementation of policy. So people can be really supportive of treatment based policy, but then not actually support building a treatment clinic in their neighborhood. And I see this playing out in my neighborhood. People do not want social services to confront this drug crisis all located in my neighborhood anymore. My neighbors are very angry about the existing concentration of those services. And people want it to be much more spread out. And so what we show in that other article is this is going to be really hard because people do not want this infrastructure near them because they see the negative externalities of say a treatment clinic being located near them.

They see people lining up out front to receive drug treatment. They see that there might be homelessness concentrated around this because of the intersecting nature of say, homelessness and drug use. They see crime starting to get located near these things because drug dealers target people who are in drug treatment. And so this leads to all these negative externalities that people don’t want. So we have a depressing takeaway from that article, which is no matter how supportive people are of policy, it’s going to be really hard to actually implement this. And so my big suggestion to policymakers knowing what we know from both of these two articles is try to frame this crisis as affecting people that are similar to the people you’re trying to convince about this infrastructure more specifically and harness some of the findings from this later article on racial identity for the pessimistic conclusions of our first article, which maybe is using race as a tool, but I’d say it’s being realistic about how people form opinions about everything in terms of policies.

So if you’re going to talk to neighbors of a potential location for a treatment clinic, talk to them about the users of that potential clinic being really similar to those neighbors. Talk to them about how this will make their lives easier to access jobs while living in a community without having to travel across the city, which is really what a lot of people face now and trying to access treatment within American cities. And so we really think these two articles play a complimentary role in trying to disentangle both how people form broad sort of vague opinions about the opioid crisis as well as specific opinions about implementation on the ground.

Matt Grossmann: So you mentioned that this is an issue where there had been a broad support for treatment across the political spectrum. There’s a lot of attention to Oregon right now reversing its very pro treatment drug policies overall. Is there a punitive turn now or coming in drug policy? Is it extending to opioids? And is there any sign that this issue is polarizing like all the others?

Justin de Benedicts-Kessner: I haven’t seen updated data. As I said before, Mike and I have been working on this in a while because of the timeline for publications. Our data is now a few years out of date, which is really unfortunate. But the current public opinion on the opioid crisis, I haven’t seen anything to suggest that more broadly, there’s this backlash to this. I think people have a lot of backlash to what I was talking about earlier, this implementation of the actual treatment policy. It’s a little bit different. And so I think we’re going to settle in a middle ground that’s still pretty treatment focused. But the real question is how do policymakers create something that everyone can agree to in terms of implementing that? I’d say there’s a lot of success in terms of the localization of implementation in other areas of policy like housing policy, which can reduce some of this backlash.

So where states can sort of tie the hands of localities and say, “We’re going to distribute this burden of the cost of treating people for opioid use disorder across the board,” similar to how say the state of California is saying that cities in California have to meet their regional housing need. I think policymakers talking to local governments about how they could actually meet the needs of substance use in their communities, but also across a state could say like, “Hey, we need to have a certain number of treatment clinics permitted or allowable in a locality based on your population or based on your overdose rates or something like that.”

And so while people may respond to the existing infrastructure to confront the opioid crisis, I think there’s a way that state and local governments can work together to make this backlash, not as evident as we’ve seen in places like Oregon. As far as partisanship, I haven’t seen this huge polarization occurring in data across time on this issue. And I think, again, part of that is just there are prominent conservatives who are still taking stances really in favor of much more treatment focused policy, in part because they are from areas where the crisis is really extreme. They have also been products of this media environment around the opioid crisis. So they’re much more sympathetic, I think, and they form those opinions. And I would hope that we continue to see policy leaders on both sides of the partisan aisle supporting these types of policy.

Matt Grossmann: There’s a lot more to learn. The science of politics is available biweekly from the Niskanen Center. I’m your host, Matt Grossmann. If you like this discussion, here are the episodes you should check out next, all links on our website. When liberals and conservatives use genetics to explain human differences, Racial stereotypes and voting for Obama and Trump, Values and racism in American immigration views, How public policy intentionally segregated American homeowners and Can Democrats design social programs that survive. Thanks to Tanika Raychaudhuri and Justin de Benedictis-Kessner for joining me. Please check out the political effects of opioid addiction frames and how the identity of substance users shapes public opinion on opioid policy. And then listen in next time.