Who believes the most "taboo" conspiracy theories? It might not be who you think

Like Henry Ford before him, Elon Musk has emerged as America’s top conspiracy spreader. But he’s hardly alone. Robert F. Kennedy Jr. is the conspiracy-theory candidate for president, and as Paul Krugman observed last summer, was attracting “support from some of the biggest names in Silicon Valley”:

Krugman didn't focus on conspiracy theory as such but on something closely related: distrust of experts and skepticism about widely accepted facts. He described this tendency as the “brain rotting drug” of reflexive contrarianism, quoting economist Adam Ozimek.

That wasn’t exactly scientific, but a new paper entitled “The Status Foundations of Conspiracy Beliefs” by Saverio Roscigno, a PhD candidate at the University of California, Irvine, is. Its most eye-catching finding is the discovery of “a cluster of graduate-degree-holding white men who display a penchant for conspiracy beliefs” that are “distinctively taboo.”

Specifically, Roscigno writes, “approximately a quarter of those who hold a graduate degree agree or strongly agree” that school shootings like those at Sandy Hook and Parkland “are false flag attacks perpetrated by the government,” which is “around twice the rate of those without graduate degrees.” Results are similar for the proposition that the number of Jews killed in the Holocaust “has been exaggerated on purpose.”

These findings are striking for many reasons. Most obviously, they go against the common belief — long supported by research — that conspiracist beliefs are more common among lower-income and less-educated individuals. They also challenge the  formulation popularized by Joseph Uscinski that "conspiracy theories are for losers," and should be understood as “alarm systems and coping mechanisms to help deal with foreign threat and domestic power centers” that “tend to resonate when groups are suffering from loss, weakness, or disunity.”

Roscigno’s findings don’t refute previous formulations so much as reframe them by adding greater nuance. For example, he finds that conspiracy beliefs are more common both among the less educated and less affluent, on the one hand, and the more educated and more affluent on the other. Secondly, he identifies the subjective group experience of threat as a key element, rather than objective “loser” status.

Even more important, his paper reveals how much more we have to learn about conspiracy theories from a rigorous social science approach. Conspiracy theory is much more mainstream, varied and ubiquitous than previously assumed, and there’s much more to be learned from studying it as an integral part of the sociological landscape. Like the recently published paper I previously covered here, this model breaks with dualistic approaches that in some sense mirror what we find troubling about conspiracism — that is, painting the world in black-and-white rather than in many shades of gray. I recently spoke with Roscigno by Zoom about his findings and where they might lead. This transcript has been edited for clarity and length.

Your paper has a dramatic finding regarding "a cluster of graduate-degree-holding white men” who tend to embrace conspiracy beliefs that are "distinctly taboo." But that's just the tip of the iceberg, because there's a whole host of questions that raises, including the role of sociology in this research, not just psychology. What led you to do the research behind this paper — what kind of questions, concerns or interests were driving you?

One of the things that motivated me was precisely the observation that sociologists hadn't really been part of the conversation. I've been interested in this topic for a while. I grew up spending a lot of time online, seeing a lot of conspiratorial stuff, having a lot of conversations with my friends about that kind of stuff. In the past couple of years, it seems like a lot of it has hit the mainstream. I remember when QAnon stuff first started fermenting online. I remember seeing posts where people were analyzing and trying to break down these “Q drops,” and sending them to my brother, like “What's what's going on here? This is something totally new.”

When I got to grad school, I thought, well, there's got to be some sociologists doing work on this. I definitely found a cluster of cultural sociologists starting to do some really interesting stuff that inspired me a lot. I also found the work of people like Joseph Uscinski and others in political science who had been doing some work and some theorization that I thought could be pulled into building a sociological approach to this.

What did you think you might learn in doing this study?

The basic question was just which groups of people tend to hold which conspiratorial beliefs. Maybe it seems like an overly basic question, but I was really struggling to find anybody in the literature that had engaged it. There's a lot of talk about who believes conspiracies in general, but there's less  attention to how different groups might be sympathetic to different claims. And I had observed in my time online that some conspiracy spaces are older or younger, in some there's more white people or more women, and I wanted to know what the variation was. That was the starting point, and then building a more sociological approach to the topic, looking at inequality and demographic variation, and then moving on to other questions.

So what did you find that confirmed that basic sociological intuition that there were significant differences, and what did you find that surprised you? 

The one pattern I really highlight is, as you said, that there's this cluster of graduate-degree-holding white men who are more favorable towards almost all the beliefs that are listed. But there are some that they are much more favorable toward, where there's a larger gap between them and those without graduate degrees. I describe these as "taboo claims." Specifically, that school shootings like those at Sandy Hook and Parkland are false-flag attacks perpetrated by the government, and the other one is that the number of Jews killed by the Nazis during World War II has been exaggerated on purpose. These two particular claims are very disproportionately held by white graduate-degree-holding men.

In addition, if you look a little bit deeper into some of the other survey research and even my own data, you can also see a concentration of medical-themed conspiracy beliefs among African-Americans, and among the less educated. Those were the two points of variation that I have been able to highlight. I suspect there are many more. But here the goal of the paper was just to demonstrate that variation exists. It wasn't to capture all of it.

That second variation is unsurprising, given that African Americans have been very ill-treated by the medical establishment. If you told most white people about the Tuskegee experiment 20 or 30 years ago, they'd think that was a conspiracy theory. But the finding about this group of more educated white men was more surprising. What have you speculated the reasons might be?

The most convincing explanation I found is that essentially this is about access dynamics. The typical theoretical focus when it comes to conspiracy beliefs tends to be toward attitudes or dispositions. I think the role of attitudes is relevant here, and I think these attitudes are fueled by a perception of threat among graduate-degree-holding white men. Maybe they see social changes that are going on, they hear how the tone of certain conversations is changing, they see how the job market is changing. So there's a perception of threat. That's where you get the attitudes.

Now the other side is the access. There's a couple of things that could be going on, but it's hard to believe that Sandy Hook was perpetrated by the government unless you've heard that claim made in some level of detail, not just seeing reporting about Alex Jones but hearing somebody really make that claim. It's even hard if you don't know what a false-flag attack is. So I suspect that graduate-degree-holding white men, particularly via online channels, are are more likely to encounter this information, more likely to run into it. We also know from scholarship on rumor that certain rumors tend to be concentrated in certain demographic networks. There's a rumor that will primarily be spread within white networks or Black networks, and that's what's going on here.

I think it's also important to take survey results about something as deep as beliefs with a grain of salt. A question I get a lot when I present this research is, “Oh, they don't really believe that, do they?” That's not really a question that a survey can necessarily answer. At the very least, we know they are checking off a survey box way more often. So if we read it with that interpretation, we can maybe say that this is kind of a transgressive act. They’re saying, “I know that I'm supposed to be checking off the other box, but I'm going to check off this one.” To me, it's demonstrating a kind of transgressive expertise, a special access to what Michael Barkun calls “stigmatized knowledge.”

So that sets off two things for me. One is the question of how you would go about digging deeper into that, testing if that's true. Related to that, it seems that survey research could be improved to ask people whether they have communicated these beliefs to others, are they deeply held beliefs that help them make sense of other things, questions like that. Have you given any thought to that?

Some of those things can be ascertained through survey research. I like the idea of asking, "Have you ever spoken to somebody about this?" or "Is this something you hold privately?" But I'm wary that survey research will give us all the answers we need. If you really want to figure out if somebody really believes something, I think you have to talk to them. You have to learn about how they live their life. You have to learn about their social relationships. It’s just like if we were studying religious beliefs. I think you have to engage at a deeper level to figure out whether that is true belief.

There's a lot of room for improvement on surveys, though. One of the biggest rooms for improvement in surveys is on the issue of prompt selection. It seems that this pattern that I noticed didn't get noticed before because nobody was asking these taboo questions on the surveys. Mostly they ask questions about COVID, and maybe a few other things. But if the prompts substantively change the findings of the survey, and nobody seems to be giving much conscious consideration about which prompts are included, there's definitely room for improvement.

You also found similar, though less dramatic, gaps between the highly educated and less educated for four other unpopular beliefs. So there's seemingly a general predisposition to conspiracy beliefs there. What other factors do you think might be involved?

I've gone back and forth, but I think there’s something I've decided on. There's this question of whether it's that they prefer unpopular [beliefs], or is it a question of, like, these things are taboo? They know these things are transgressive, they know these things violate a deep social norm. I'm pretty sure it's the taboo.

But this can be pretty easily tested. There are beliefs that are very unpopular but are not particularly taboo. If you ran a survey that included something like belief in a flat earth, if I'm right we wouldn't expect white grad-degree men to be high on that. When I say “taboo,” I basically mean that if you said something like this in public you would face some kind of social sanction. If I told my co-workers that I thought the earth was flat, they might laugh at me. If I told my co-workers that I thought the Holocaust was exaggerated, it would be a very different story.

Do you have some thoughts on what research you might be able to do to make more sense of this? 

One thing that could be done is looking at a really wide variety of prompts and seeing what kinds of patterns are going on. In this one, I'm working with 15 claims and trying to draw a common thread. If you worked with a much larger set of prompts — I know some of those data sets exist — I think it would let you articulate that a little bit more clearly.

But that's only one way to approach the issue of typology. You could start at the point of "there's a group of people that tends to hold these beliefs," and try to describe those particular claims. You could also start by looking at the claims and trying to find narrative threads between them. You could also define the claims by the relation to some authority, which is kind of what I'm doing with the taboo stuff. So I'm not quite sure how to address that yet.

You also found roughly equivalent subsets of respondents who held both of those claims [about school shootings and the Holocaust] and who disagreed with both, providing a convenient comparison. They differed in terms of extremism and social media use. So what can you say about those differences and how they interrelated?

I already mentioned the question of access. I think social media use gets at that access question. Those who agree report higher levels of social media use by every platform, particularly by anonymous image boards like 4chan and 8chan. So at the very least, if we think about people stumbling into these beliefs kind of accidentally, if you're on 4chan more often you're a lot more likely to run into one of these. In addition, there's some interesting work being done on information-seeking strategies online, and some sociologists have pointed out that people with different social positions have different strategies that may lead to different results. So a possibility relative to social media is that white men with graduate degrees, when they're doing research, the steps they’re taking may be different from some other groups, so they're more likely to end up at a certain point.

Relative to political extremism, that's a bit more complicated. There's definitely some exciting research that's going on about radical political beliefs and their relationship to conspiracy beliefs. Something I want to point out is that the white grad-degree men who agree are way more on the political edges, which maybe is to be expected. They're identifying as extremely liberal or extremely conservative way more often. We get this U-shape. These two taboo claims, at least according to this measure, are not right-wing phenomena. There is a big cluster of people that identify as very liberal and agree with these things as well. I suspect this measure isn't picking up on everything it could be. In the time that I've spent in politically radical spaces online and within the conspiracy milieu, the way people identify politically — there's a lot of variety to it, and “liberal” and “conservative” descriptors may not resonate with a lot of these people. But at the very least we know that people on the political fringes tend to be more charitable towards these claims.

Over time, erosion of social trust seems to be related to a rise in conspiratorial beliefs. It would make sense, just in terms of people who feel skeptical of the existing system, for that to show up more, regardless of whether they are left-wing or right-wing. Do you have any thoughts about that?

This is something else I think that sociologists have to bring to the table: What's with the structural context of these situations? There's some evidence that countries with higher levels of social inequality, higher levels of corruption, tend to demonstrate or report higher rates of these beliefs. We know that it's tied to structural conditions. The collapse of institutional trust is a huge piece of this. If you look at graphs of trust in the federal government over time, or trust in the press over time, they're really at historic lows.

That has to play some role. Because when we talk about conspiracy beliefs, in the simple definition we’re talking about claims of elites doing something in private, but we’re also talking about something that counters the official narrative. So in a situation where historically few people trust the producers of the official narratives — in part the government, in part the press — we would expect people to be more doubtful of those things.

But when we talk about social trust, I don't necessarily think belief in conspiracies means a low level of social trust in general. I think it means a low level of trust in particular institutions. But in order to believe a conspiracy you have to hear it. It's probably from somebody you know, and you have to trust them when they tell you that. There's a rumor scholar, Gary Fine, who says that when trust in institutions is questioned, trust in informal networks is revealed. So there is a social trust that exists. It's much more decentralized. It's not in a particular institution and it’s social trust, rather than institutional.

One thing your paper suggested to me was looking at how beliefs in conspiracy theories co-vary, meaning what beliefs go together or tend to negate each other, and how that might change across status lines. I was specifically interested in those white male graduate-degree holders. Are there any beliefs that they accept less than other people? Do you have enough data to look at that yet? 

I think enough data exists that we can probably answer that, but I don't know for sure. In this particular data set, there are none that they were less likely to believe in. For very mainstream beliefs — the idea that “one percent” of economic elites control the government and economy, the idea that Jeffrey Epstein was murdered — these are beliefs held by 50% of the general population and also held by about 50% of white men with graduate degrees. In this data there wasn't a single belief that these white male graduate-degree holders were less interested in. That was stunning to me. I was actually very surprised by that.

Want a daily wrap-up of all the news and commentary Salon has to offer? Subscribe to our morning newsletter, Crash Course.

But it's possible. We have these two types that I'm describing, the medical ones and the kind of taboo ones. It's possible there may be some medical ones that white men with graduate degrees are less likely to agree with. But it is hard to say, because this data clearly suggests that graduate degree holders are more into all of these claims.

We spoke earlier about “prompt selection” and things that perhaps aren't being asked about. Do you have anything specific in mind?

There's a lot of things that aren't being asked, definitely more than are being asked.  At least in this paper, my starting point is the simple definition of the conspiracy belief which is, again, basically that a group of elites are plotting something in private. If that's our conceptualization, then the universe of possible things to ask about is massive.

For instance, if that's our conceptualization, why don't we ever ask about institutionally verified conspiracies? For instance, Watergate fits that definition just fine, COINTELPRO fits that definition just fine, Tuskegee fits that definition just fine. To me there seems to be a mismatch, an unacknowledged element to the definition, which is that it has to counter some official narratives. But even if we include that second part in conceptualization, there’s still tons and tons of stuff.

I read a very interesting paper this past week. This came out as a content analysis of TikTok, but it’s specifically about the conspiracy theory that Taylor Swift is secretly gay, and she's closeted and dropping all these hints in her tracks. Maybe people will say that isn't a politically consequential conspiracy theory, but it’s within the realm of conspiracy claims by any definition.

I've even heard that and I don't follow Taylor Swift news at all. It's clearly out there. 

Yeah, if I had to guess, if you polled the demographics it would disproportionately be women. So that makes me think, OK, a lot of studies emphasize that men are more into this stuff. Does that have something to do with the prompts that are selected for the surveys? How does that come into play?

I bring up the Taylor Swift example to demonstrate that the realm of  things under this blanket is, like, so large that trying to generalize any kind of research findings to the entire world of claims about elites doing sneaky stuff ends up being very difficult. I suspect there are claims that graduate-degree holders are more into that we haven't quite figured out yet. I suspect there are claims that women are more into that we haven't really figured out. I'd like to see a lot more, a) alignment between the conceptualization and operationalization and b) experimentation within that. We have a big world of things that fit this conceptual framework.

We’ve talked a bit about “collective identity” as a useful concept and you've said “it applies to all varieties of conspiracy cultures." Could you expand on that?

To be totally sociological, collective identity is useful in understanding all kinds of cultures more generally. Within conspiracy cultures, there's a couple things going on. If we talk about rumors, if I tell you some finding before it's published, it feels like you're in the know, it feels like you have a piece of secret information. It's exciting, it feels good. It also creates a bond between people that I think can be part of identity. So that's one level.

There's also the level that gets to the question of institutional distrust. There's a general sense in this country that, you know, people like us — whatever “us” means — are being screwed over by elites in some faraway place. We can't really see what's happening over there, we're not in the rooms where these decisions are made. I think there's a very general sense of that. And who “people like us” ends up being defined by is, I think, very important, because different people are going to understand it in different ways. There’s a general sense that there's opaque power that's screwing us. We don't really know where it is, or what's happening. You hear that kind of sentiment a lot in this milieu.

There's also collective identity more overtly. If I make the claim that white people are being replaced in this country — which to me is one of the more consequential conspiracy claims — I'm invoking a very specific identity, saying, “Hey, we collectively are under threat and need to do something about it!” So some conspiracy claims, even in the claim themselves, name the in-group or name the out-group. It will say who the “we” are, who the mysterious “they” is. Identity plays a key role there as well.

What stands out for you as the next steps? What questions need answering that follow from what you've done so far? 

To me, a lot of the most interesting questions are about how, when and why these beliefs matter, which I do think are better suited to qualitative methods. There's been very little in the way of qualitative inquiry into conspiracy cultures, with the big exception of Jaron Harambam, whose work has been very inspiring to me. Back to this matter of collective identity, something he points out that I find intriguing is that there are all kinds of conflicts over identity, even within the conspiracy milieu. There are people who understand themselves as aiming to get new converts to the movement, and other people who understand themselves as basically having given up and clocked out. There's all kinds of variation within the community.

Also, part of my reasoning for wanting to do qualitative research is that I feel like a lot of people in the conspiracy milieu feel like they're being studied from afar by people that aren't talking to them at all. I think that absolutely adds to the resentment.

I saw a tweet recently from somebody who is loosely in these circles that basically said, “How come none of these conspiracy theory experts are even talking to us?” If you are an expert in Amish culture, you'd probably want to spend a lot of time talking to Amish people. If I were studying the student movements that are going on right now, I'd probably be down at the encampments hanging out. It's not like believers in conspiracies are a small or fringe minority group that's super-hard to access. Some of these claims are totally mainstream, and even for the more taboo ones that you might envision would be hard to do qualitative research into, they're concentrated among graduate-degree holders. So in some sense those of us in academia are exceptionally well positioned to engage these communities at a closer level. So I definitely would like to do qualitative research in the coming years.

Finally, what's the most important question I haven’t asked? And what's the answer? 

I can tell you a question that I get whenever I present my research to my undergraduates, but I'm not going to answer it. I give this whole presentation and at the end they’re like, “What are the ones that you believe in?” That's not my role as a sociologist. [Laughs.] So that’s my favorite question.