My Columbia political science colleague Eunji Kim shared with me her new book on American public opinion. In its overall structure, it’s the most well-written and coherent academic political science book I can recall reading. Much better organized than our Red State Blue State book, which had a general theme and lots of little studies but no coherent message, and various other political science books I’ve read that are either a collection of only loosely-related articles, or, at the other extreme, a single article puffed up to book length. I enjoyed reading it from beginning to end. It’s kinda too bad because almost nobody reads academic books from beginning to end, but at least this one reader appreciated the care that was put into the book, not just the research and writing but also the structuring of the argument.

I have a few thoughts and questions:

1. The political science consensus

Kim present herself as going against the political science consensus by studying entertainment media. And I guess she’s right; Figure 2 of the book seems pretty clear.

Still, at times I feel she overstates the case. On page 3, she writes, “We tend to think of our political beliefs as well reasoned and carefully considered, or, at worst, determined by what’s happening around us right now. In this hyper-politicized world full of partisan news media, it seems implausible that something as frivolous as the latest reality TV show from Netflix, cop shows, and superhero movies, to name just a few, could possibly affect something as profound as people’s political attitudes.”

And then she continues with, “This is a reasonable viewpoint shared by the vast majority of political observers and scholars of public opinion.” What’s the evidence for this claim? I’m not trying to be picky here! I would guess that most political scientists do think that entertainment could affect political and social attitudes. Indeed, entertainment news often includes speculations about political effects of TV shows, celebrity endorsements, actions on the sports field, etc. Granted, these entertainment news stories are not written by political scientists, but political scientists are aware of such speculations–I know I am! One problem here is that lots of people, political scientists and others alike, tend to overestimate the effects of things that come to mind (this is related to Tversky and Kahneman’s “availability bias”). For example, when Joe Rogan endorsed Donald Trump and Taylor Swift endorsed Kamala Harris, there was a lot of talk about what effect these might have. I don’t think anybody thinks these effects are zero; at the same time, such effects of individual endorsers are not going to be huge, because there are a lot of celebrities out there.

Speaking more generally about entertainment’s effect on culture and political attitudes, again, there’s been lots of discussion and speculation on the effects of cultural influences including movies (Dirty Harry, Death Wish, Star Wars, Rambo, Back to the Future, Do the Right Thing, The Matrix, etc etc etc), business stories (McDonalds, De Lorean, Microsoft, Apple, Google, Tesla, etc.), pop stars, sports stars. You get lots of concern on the right about permissive cultural messages encouraging permissive attitudes and liberal voting, and corresponding concern on the left about vigilante movies and corporate success stories encouraging conservative social and political attitudes. I’ll take Kim’s word for it that political scientists have published very little on the topic, but I don’t think that means that they (we!) find it “implausible that something as frivolous . . . could possibly affect something as profound as people’s political attitudes.” I think it’s more what she talks about later in the book, that these things are very hard to study! We’ve pretty much ceded the field to sociologists. So, fair enough to say that this is an under-researched topic; I just think she’s setting up a straw man by saying that “the vast majority of political observers and scholars of public opinion” that entertainment media doesn’t matter.

Also, on p.139 Kim writes, “A credible possibility exists that entertainment media may still impact policy attitudes,” and then she cite some studies. Some of these studies may be iffy–I’ll get back to that later–but, in any case, doesn’t this kind of undercut her claim that nobody studies these things? But, yeah, if it’s really true that “80,000-plus academic treatises” mention Fox News (as she says on p.177), then fair enough to make the point that the effects of entertainment media are understudied.

Umm, here’s another place where I think Kim slightly misrepresents the field of political science. On p.179, she quotes Pye and Verba (1965) as saying, “Some would say that politics may be found everywhere–in the club room and the business office, among schoolmen and churchmen, and even the household–but surely politics assumes its great dimensions only when its stage is the state and its powers can shape the law of the land.” She then writes, “While many intuitively or instinctively postulate that politics only reveals itself in the corridors of power, I humbly diverge from that perspective. I align myself with those who recognize politics as a perennial undercurrent, coursing through the veins of everyday American life.” But Pye and Verba weren’t saying that politics only reveals itself in the corridors of power. They explicitly noted that politics occurs everywhere. What they said is that politics assumes its great dimensions” only with affairs of state. So I again I Kim is arguing against a strawman.

On the other hand, maybe the narrowness of political science that she’s arguing against is not such a strawman. Yes, political scientists recognize the political aspects of interactions at work, at home, on the football field, wherever–but if you look at what’s published in political science journals (including by me!) it’s pretty much all about government policies, political parties, and public opinion, maybe with the occasional article about labor union elections or church politics or whatever. If you want to look for articles on politics within firms, within families, etc., you’ll have to go to the economics and sociology literature. And that ain’t right! The insights of political science should be relevant to understanding small-scale politics, in the same way that economics and sociology provide insights in the small as well as the large.

2. Upward mobility and economic pessimism

Kim argues that entertainment media portray an unrealistic world of just deserts in which anyone who is deserving and works hard should be able to achieve success. How does this fit into another thing we hear about, which is a general pessimism in America, and indeed around the world. This came up during the 2024 campaign: by the usual measures, the economy was going well–not booming, but going fine, with low unemployment, moderate and decreasing inflation, slow but positive GDP growth, etc.–but there was a general sense of there being a bad economy, and also a larger sense that America was no longer operating as it should, that traditional social and economic roles were disappearing, etc. This was taken to be a big part of the vote swing to the Republicans.

I’m not saying there’s a contradiction here–it’s possible for people to feel that America is still falling apart while still being cheered by reality-TV success stories. I’m just not sure how this all fits together, and I didn’t see it addressed in Kim’s book.

Let me put it this way: the entertainment media have given people the sense that with hard work anyone can climb the ladder, and this can push people toward supporting traditional Republican low-tax, low-spending policies and opposing redistributionary policies of the Democrats–but at the same time people seem to feel that their economic situation is precarious. Again, this is easy enough to explain in traditional political-science language: the voters feel that times are tough for them and so they oppose government giveaways to the undeserving others, and you could say that exposure to “Cops” gives people the sense that the poors are bad people and that exposure to “Shark tank” gives people the sense that poor people have only themselves to blame–but people are not applying this logic to themselves, right? When it comes to their own economic problems (the price of eggs or the disappearing industrial base or whatever), then people see themselves as the victims. So there still are some pieces of the puzzle that I’m not understanding.

3. What is “meritocracy”?

Kim mentions meritocracy many times in the book but I didn’t see a definition. It seems to me that meritocracy has two meanings. The first is simply that people’s station in their jobs and society is based on their abilities or merit. The second is that people with merit get to run things. I think the second definition is more accurate: the idea that the people with merit get to run things is the “ocracy” part of meritocracy (“government or the holding of power by people selected on the basis of their ability”). In that case, though, the concept of meritocracy is itself self-destroying, because one thing the people with merit will do with their power is to get favored positions for their friends and family. I elaborate on this point here and also here. I think this issue is important because it moves the debate beyond “Is meritocracy good or bad?” to “Can meritocracy exist at all?”

To put it in political science terms: You have a population whose attitudes on social and economic mobility are being affected by entertainment media. But then there’s the question of how the attitudes transform to votes. I’m not saying Kim needs to figure it all out in this book; I’m just saying that, with “meritocracy” itself having such a slippery definition, this makes it even harder to interpret the book’s empirical results in terms of their impact on political outcomes–this is the issue I raised in point #2 above.

4. Reliance on shaky studies by others

In various places in Kim’s book she refers to various studies that I don’t trust. These are a bunch of papers published roughly around the notorious 2010-2015 period when the replication crisis in social science was at its height. I wasn’t quite sure what Kim’s take was on these studies. On one hand, she seem to be taking them at face value, which I think is a mistake, as I see these studies essentially as noise-mining exercises. On the other hand, she also downplays the importance of those results (or, I would say, those unsubstantiated claims).

Here are some examples:
pp.6-7. “It turns out that those who vote in schools are notably more inclined to support school funding initiatives due to ‘contextual priming.'” I looked up this paper and it’s based on observational data from one referendum in one year in one state. Indeed, on page 48 of Kim’s book she does a good job of pointing out the weaknesses of this sort of one-off observational study.
p.7. “Even chance encounters, like observing someone in apparent poverty in the streets of Boston, reportedly reshape attitudes on wealth redistribution.” This was a notoriously noisy experiment where the result went in the opposite direction as expected and did not even reach the conventional level of statistical significance; see here. My point is not that we should only talk about “statistically significant” results but rather that this study could easily–indeed, I think would usually–have been presented as a null finding.
p.7. “Scholars even found evidence of how shark attacks or lousy weather affected people’s voting behaviors.” This shark attack study has major problems, indeed I think it would be fair to describe it as debunked; see here and here.

Kim follows up with, “much more systematic than the political impacts of shark attacks or random encounters with out-group members would be the influence of entertainment media, given the ubiquity in our daily lives.” I agree. That’s one reason I think it’s funny that she reported those questionable studies with a straight face! Indeed, my colleagues and I have argued that those silly claims of large and persistent effects from trivial inputs cannot be all or mostly true for mathematical reasons that a large number of large and persistent effects cannot coexist:

Kim continues by saying, “Perhaps the glaring irony is that, while political scientists have no trouble believing the powerful impact of ostensibly arbitrary or seemingly irrelevant events on public opinion, they have been reluctant to study the media that citizens primarily consume.” To that I reply: (a) lots of political scientists don’t believe those claims about the purportedly large and systematic effects of shark attacks, stranger encounters, etc. Yes, some do (Achen and Bartels are decorated political science academics, and that streets-of-Boston-study did win two awards), but lots of political scientists (including Anthony Fowler, Andy Hall, and me!) are skeptical and think that the social priming research is a dead end; and (b) As Kim explain so well in her book, it’s hard to study the political effects of entertainment media! I don’t think political scientists are reluctant to study entertainment media; it just takes a lot of effort so we gravitate to where the data and research questions are cleaner. Indeed, one of the big problems with social priming research is that it’s easy to run an experiment or grab observational data and perform what appears to be a well-identified statistical analysis. So Kim is on to something here . . .

To continue:
p.43. “The emotional resonance or ‘affective imprint’ of these media narratives can leave a lasting impact on our attitudes and behavior, often occurring at an unconscious level.” I’m very skeptical of these claims of large and persistent unconscious effects. I looked at one such claim carefully and it disintegrated under careful inspection; see here. I don’t these claims make a lot of sense, also I don’t find the evidence given in support of those claims to be at all convincing. On the plus side, I think Kim’s main story is strong on its own–for one thing, there’s no need to believe that effect of Shark Tank etc. on attitudes about meritocracy etc. are “occurring at an unconscious level”–so I think she could strengthen her argument by detaching it from the questionable replication-crisis-era research that she is citing.
p.49: Kim skeptically discusses a paper about Harry Potter that “explores the hypothesis that avid Potterheads might be less inclined to support Trump.” I think that paper might have ben a parody; in any case, it has problems of causal identification (see here). In any case, I think Kim’s discussion of that paper was good, in that she points out the lack of any coherent theoretical story or model there.
p.139. “MTV’s 16 and Pregnant is credited with altering rates of teen childbearing.” I haven’t looked into this myself, but that claim has been contested.

5. Media consumption

This is something I know very little about. I learned a lot from all the examples in the book. There was one place where I can share something. On page 34, Kim writes, “Remember how sports fans run for cover when their team’s performance is too depressing to watch? Politics, just like any team sport, works similarly.” I wrote a paper about this in 2016 with Doug Rivers, David Rothschild, and Sharad Goel. We found strong evidence for differential nonresponse. This wasn’t media consumption, it was response to opinion polls, but it seems related. Rothschild and I also made the point, with additional data, in an article that year in Slate.

On page 77, Kim reports that 40% of respondents said they watched America’s Got Talent, 31% watched Shark Tank, and 27% watched Hell’s Kitchen. Elsewhere she writes that this is the percentage of people who described themselves as “regular viewers” of the shows. Wow–those are huge numbers! I say this as someone who’s never watched any of the shows on the list of Figure 3.7, except that I watched part of an episode of Survivor once, about 25 years ago. So I’m not tuned in to the viewing habits of the average American. Fair enough! But I also wonder what people mean when they say they are regular viewers” of a show. On p.76, it says that American Idol and America’s Got Talent had average audiences of 5-10 million. 10 million is 3% of Americans. Can it really work out that 3% of people watch the show at any given time, but 30% are regular viewers? What does it mean to be a regular viewer if you watch the show less than 10% of the time? I’m not saying Kim is wrong here; I’m just having difficulty making sense of these numbers. It’s not like in the 1970s, when popular TV shows had ratings in the 30s.

6. Conflicting messages from entertainment media and social media

As discussed above, Kim argues that entertainment media has been pushing an if-you-work-hard-in-America-you-can-make-it story, what she calls “meritocracy” and that this helps to explain why many Americans oppose policies of economic redistribution from rich to poor. Of course this isn’t the only reasons for people to oppose redistribution–there are economic arguments, political ideology, partisanship, all sorts of reasons–; she’s just saying it’s part of the story, and a shift of even a few percentage points can make a difference. And I can buy her argument about the importance of entertainment media.

But what about social media? It seems like Facebook, Twitter, etc., are full of people pushing get-rich-quick schemes, people making a million dollars off of some scam or another. These don’t represent hard work or meritocracy; they represent a way that people are getting something for nothing. You might see this and envy those people, or want to be them, but I wouldn’t think that a natural reaction to seeing social media posts on house-flippers or whatever would be to conclude that the American economic system is fair.

So here’s my concern. Kim is pitting two economic narratives against each other:

A. The economic system is complicated; we need regulation and transfer payments to keep things running effectively. I’d call this the liberal or Keynesian view, the idea that for moral reasons the government should work to reduce income inequality, also this will be good for the economy as a whole, by increasing employment and output.

B. People pretty much get their economic just deserts; if you work hard you can make it in our society. I’d call this the conservative or monetarist view, the idea that transfer payments are a moral hazard discouraging people from working, and that limited government will grow the economy and help everyone.

Kim doesn’t quite say it like this, but I get the impression that she’d associate narrative A with the sorts of economic statistics you’d see in the news media (GDP growth, unemployment rate, inflation rate, etc.), while narrative B is being pushed by social media.

But what about this other narrative:

C. The system is rigged: no matter how hard you work, you’re just spinning your wheels; the real money is being made by people who are doing some scam or who’ve found their way into some passive investments (real estate, bitcoin, whatever). This is a cynical narrative–I guess it has some truth; it’s still cynical–and it’s different from A and B above.

My impression is that narrative C, which can have liberal or conservative policy implications, is big on social media. And I’m not quite sure how it fits into Kim’s book, which is so focused on the just-deserts message being pushed on entertainment media. Even if a tiktok account is telling you that anyone can become rich through bitcoin, that’s not a message of “meritocracy” as it doesn’t connect to hard work or to any sense of personal merit.

I sent the above to Eunji and she replied:

To address your first point regarding the political science consensus, I understand your concern about overstating the skepticism around entertainment media’s role in shaping political attitudes. As you rightly point out, discussions about the potential political impact of celebrities, TV shows, and movies do occur regularly in the media and among political scientists, though they’re rarely the subject of academic writing. My intention wasn’t to suggest that scholars entirely dismiss these effects, but rather to highlight how understudied they are compared to more traditional political research. In my effort to write a more accessible book, I see now that my argument might have come across as setting up a straw man.

That said, I think part of the reason my writing leans in this direction is due to my own experience in the field. I can’t help but reflect on the immense skepticism I faced throughout grad school regarding my focus on entertainment media. Well, part of the skepticism, I think, arose from the fact that I am not American–a ridiculous point of view that I unfortunately heard repeatedly. It often felt like my work wasn’t being taken seriously, and I was told repeatedly that it wasn’t “real” political science. Many scholars have cast serious doubt on whether entertainment media can truly shape political attitudes, and that created a sense of swimming against a very strong current in my research. Many of them have all been generous mentors, and I admire their scholarship, but as a grad student, it felt a bit daunting that my argument goes against what some of the towering figures in my subfield wrote.

One way I think about contemporary economic pessimism, which I touch on in the book, perhaps more implicitly, is the distinction between how people perceive their personal situation (egotropic) versus how they think about the national economy (sociotropic). Many Americans report feeling a heightened sense of personal economic insecurity, but they also continue to believe that anyone can succeed if they work hard. What’s interesting is that those who feel economically insecure on a personal level are often the ones most drawn to rags-to-riches stories in entertainment media. This may partly explain why the perception of personal hardship coexists with the belief in upward mobility.

As for the 2024 election, I think the core intuition of my book does help explain what happened. Throughout the campaign, liberal and mainstream media were telling Americans that the economy was doing better than they thought, but on platforms like TikTok, there were millions of videos where everyday Americans complained about the rising cost of basic items, like burgers at McDonald’s. Here, we see a significant gap between the elite/mainstream media narrative and the media consumed by ordinary Americans–those TikTok videos, in particular, shaped economic perceptions much more powerfully for a wider swath of the electorate. In a way, the entertainment and social media content that reflects economic frustrations could have been far more influential in shaping perceptions of the economy than the messages coming from traditional outlets. TikTok and social media platforms are obviously very different from reality TV I studied in this book, and my current working projects are dealing with these newer topics.

Your thoughts on the definition of “meritocracy” are very helpful. This is something I’ve personally struggled with, as the definition remains so fuzzy, and every scholar seems to use it in different ways. I’m not sure I fully agree with what James Flynn argues–that to the extent people with merit get higher status, they would use that status for nepotism, so to speak. In real life, of course, that’s often what happens, but I always find it interesting that in America, job referrals are so widely accepted. They happen far more frequently than in places like Korea, for example, where most systems are based on standardized exams and applications. When I think of meritocracy, I tend to think of the common, though admittedly fuzzy, definition that many historians and sociologists seem to use—namely, the belief that anyone who works hard can get ahead economically and socially in America. I think this definition makes even more sense when considering the downstream consequences I showed in Chapter 6, where people begin to view the rich as more deserving, attributing their success to internal factors. The electoral implications of meritocratic beliefs (whether believing in meritocracy makes you more likely to vote for a conservative party) are much harder to prove, and honestly, that’s something I didn’t even attempt in this book. The case of Trump, however, is an interesting one because he is essentially a product of reality TV. That was an opportunity I grabbed to talk about the electoral consequences of entertainment media in the last chapter and in my separate paper published in APSR.

On the studies you flagged as shaky, I completely understand your concerns and appreciate your rigorous eye. The studies you mentioned do indeed come from that fraught period of social science (!), and I can see now how they might undercut the strength of my argument. I will move away from relying too heavily on them, especially in light of the replication crisis. I want the core of my argument to stand on more solid ground, and your suggestions will definitely help me revise the evidence base.

On the question of media consumption and the numbers behind “regular viewership”: your QJPS paper on the mythical swing voter is actually what inspired Jin Woo and me to write a paper about temporal variations in partisan media consumption (we even cite your paper in that one)! For the percentages of people who say they regularly watched shows like America’s Got Talent, 40% is the figure that respondents report. However, the definition of “regular viewer” is, of course, extremely fuzzy, and this is a long-standing debate among media scholars about how on earth we can reliably measure these things. Surveys tend to inflate these numbers, where people’s perceptions of “regularity” vary significantly. I used both survey and behavioral data from Nielsen, and I see how that could lead to confusion. When survey respondents report that they regularly watch a show, they might include people who watch it on-demand, via streaming, or in a delayed fashion, which inflates the number of self-identified “regular viewers.”

Lastly, regarding the conflicting narratives and messages from entertainment media and social media: you’re right that countless different narratives are circulating on social media. One interesting observation, though, is that many young people now view social media creators as aspirational career models. What’s particularly striking is that becoming a successful social media creator who generates millions of views (and, consequently, significant income) seems to require little more than sheer talent. Unlike traditional career paths, which often demand family money or connections, all you need is a phone and some compelling content. I often wonder how this shifts the way younger generations think about meritocracy and who is truly deserving of success. In this economy, it seems that if you’re poor, you have no excuse–after all, you’ve got a phone too!

So that last point is that, when some Youtube or Tiktok video is promoting some get-rich-through-passive-income scheme, what they’re implicitly selling is not the scam that they’re promoting but rather the meta-message provided by their success: it’s not that you, the viewer, can or should become rich by flipping houses or selling Beanie Babies or whatever, but rather that you can become successful by creating your own Youtube channel. The meritocracy, or whatever it is, comes from the idea that anyone with a phone can, by working hard enough, become a successful influencer.