Why You’re Fooling Yourself About Fooling Yourself About Fake News

Josh Marshall, who is generally one of the better political commentators out there, recently wrote a piece called Why You’re Fooling Yourself About Fake News. The point of the piece is that liberals who believe that fake news sways elections are wrong. Fake news, says Marshall, is a demand-driven phenomena: the people reading and believing fake news are reading it as an expression of beliefs that are already set. This has been common wisdom for a while, and during the flurry of calls from reporters I got after the election it was probably the top question I was asked — Sure, people saw that Hillary may have killed an FBI agent — but given how extreme you’d have be to believe that, how could it sway any votes?

So caveats first. A lot of what we repost falls into what I’ve called “Identity Headlines“. I repost something on the homelessness problem in Silicon Valley partially to raise awareness, but largely to show that I’m the sort of person who cares about the homeless and hates Silicon Valley dicks. I don’t even need to read the article to achieve this effect, and if I’m like most people on most days, I probably don’t.


Fake news is similar, and in this way is demand -driven. A lot of people really, really, really hated Hillary Clinton, but after multiple hacks and leaks and congressional investigations the actual world was still not producing headlines that adequately expressed the level of hatred and suspicion many people felt. So the market stepped in and created some fake ones. Headline-as-bumper-sticker problem solved!


Now, that headline above gets much closer to how a lot of people felt about Clinton than all of the stuff about how she may have said that some Bernie supporters were living at home with their parents or that people were deplorable. So people share and repost it, and it’s doubtful that the people who did that, at least initially, changed their vote. As Josh says, they were in the market for this from the start.

But I can’t stress this enough: this is only one half of the equation, and I remain confused that smart people can’t see that.

Fake News Headlines Had an Impact on Beliefs of Non-Posting Readers

The first problem with the “fake news only solidified partisanship” is what data we have about readers (not re-posters) tends to contradict that. Actually, “contradicts” is an understatement.

Now caveats on the analysis I’m about to show. Surveys on “Did you hear about X?” have some issues. Some people, put on the spot about what they know, want to appear knowledgeable, and a lot of people on opinion polls want to express a sort of personal orientation to the world more than a specific reaction to poll questions. There’s a survey on conspiracy theories, for example, where the pollsters asked about a fictional conspiracy — the North Dakota Crash — only to find that 32% endorsed it. By fictional conspiracy, by the way, I don’t mean just “conspiracy that is not true” but “conspiracy the pollster invented that has never been talked about before.” In a similar vein, we’ve never quite been able to tease out how many Americans believe that Obama is a Muslim and how many of them have been using that question as a proxy for how they feel about Obama.

But caveats aside, the one poll I’ve seen that’s actually asked folks what stories they saw and if they believed them is this one from Buzzfeed. And even when mentally accounting for some general-issue polll weirdness the results are still surprising.

The first finding was that a large percentage of people had seen fake headlines — as large as any real story. And the exposure was far less polarized than one would think. For example, for the infamous FBI Murder-Suicide headline, 22% of respondents had seen it, but about 40% of those who had seen it were Clinton voters.

Now we know that the Clinton voters who saw such anti-Clinton and pro-Trump fake stories are going to be much more likely to reject those stories as false. But how much more likely? Keep in mind these stories intimate that Clinton had an FBI agent killed and that the Pope endorsed Donald Trump. They are not subtle.

What we find surprised even me. Among Clinton voters that remembered a fake headline, the majority believed it in almost all instances.


That FBI agent story, for example? Over 50% of Clinton voters who remembered it believed it. Over 60% of Clinton voters who remembered the story about Trump protesters being paid believed that as well.

Is it a perfect way to assess this? No — I have some issues with the methodology of the poll. I think a different approach would shave some points off here. But even if you cut these effect in half, they are still completely inconsistent with the idea that fake headlines don’t influence belief, or that partisan leanings innoculate people against them.

Don’t like the study? Do another one (seriously, do it — we need a better one). But at the moment it’s the best evidence we have as to what went on, and it supports the idea that headlines matter.

If a bunch of Democrats went to the polls to vote for Hillary despite the fact she may have murdered an FBI agent, is it hard to believe that at least some Democrats didn’t stay home out of disgust? In fact, isn’t it probable?

How many? We don’t know. But we do know that it would only take about a football stadium’s worth of voters across three states to sway the election. This poll suggests that millions of Clinton voters went to the polls believing damaging stories about Clinton that were completely false. Millions. It really does not seem a stretch to think this had an impact.

What do these voters look like? I’ve met them, personally. Folks — Democrats, even — that believe that maybe Seth Rich, a DNC worker who died in a street robbery last summer, might have been killed because he knew too much about the DNC leaks, or was going to leak additional material. When pressed, they retreat into “well, I didn’t say definitely, but something is weird there.”

Um, no. Nothing is weird there at all. What’s weird is seriously considering this possibility.

The Weird Nihilism of the Fake News Didn’t Matter Argument

So we get then to the second point: maybe people did believe these stories, but they didn’t have an effect. Clinton voters thought — oh, the FBI agent died suspiciously, but that doesn’t prove anything. And there’s a 5% chance that Clinton may have killed a 20-something computer-voting specialist at the DNC, but that’s OK, because I like her policies.

Taken to it’s logical conclusion this amounts to a sort of nihilism, because it implies that “real news” doesn’t matter either.

Josh believes, for example, that the Comey letters did serious damage to Clinton, and may have lost her the election (for what it’s worth, I agree). But if we believe that people are innoculated against headlines by partisan belief, how is that possible? Under the pure “Identity Headlines” analysis no one’s opinions should have changed — pro-Clinton folks would read the Comey letter headlines and think “Well, more interference” and pro-Trump folks would read the headlines and say “Yep, Crooked Hillary” and the needle wouldn’t move.

But of course that’s not what happened. And I’ll quote Josh here:

Just to put my cards on the table, I believe there is a good likelihood, probably even a probability, that if the Russian subversion campaign had never happened and James Comey had never released his letter, Hillary Clinton would be prepping to become our new President. My own guess is that Comey’s letter had the bigger impact. These were both profoundly damaging events in the race and Clinton lost by very tight margins in most of the newly (hopefully temporarily) red states. I see little way to challenge this assertion.

So the Comey headlines hurt, but the fake headlines did not? How is this a tenable position?

Not to belabor the point, but imagine a world where Comey did not release a letter about the Weiner emails, but a large percentage of voters believed he had, due to fake headlines. Why would we expect that the real news would have a different impact than fake news? To a gullible reader, the fake event and the real event have equal impact.We don’t live in a world where we experience fake news differently from real news — to the reader they are equally mediated events, and indistinguishable from one another.

There’s a good argument to be made that events, and campaigns, and news, and scandals don’t matter for the vast majority of voters. They probably don’t. But to say that fake news can’t swing an election you have to move from an argument that people are resistant to news to one that they are impervious to it. And this is in fact to say that nothing in a campaign matters — not positions, not actions, not scandals, not ill-considered comments. Not the campaign itself, or the coverage it generates. That just doesn’t make sense.

9 thoughts on “Why You’re Fooling Yourself About Fooling Yourself About Fake News

  1. Hey Mike. I wrote an article a while ago (it hasn’t been published yet and I wrote it before I heard of Josh’s article, which I will now read) entitled “Fake News isn’t Your Main Problem” and it goes waaaaaayyy beyond this. The real problem isn’t even fake news. Fake news may be why Hillary lots by margins. Some bigger problem is happening that Donald Trump even made it as the Republican candidate in the first place. Regardless who won the election, America is in trouble. The world is in trouble. And while Facebook isn’t blameless or neutral in all of this, the problem is much deeper than that imho. I will share my article when it’s out. I expect rebuttals from everyone 😉

    • I would say that Fake News isn’t the deeper problem as well — the real problem is a collapse of authoritative institutions that people (and particularly Republicans) trust to mediate and settle questions of truth. Institutions — e.g. the scientific community, or the press, or the universities — are, in general, how we settle questions of what’s true.

      A twenty year crusade against the “liberal press” has left Republicans without an institution capable of ruling on questions of fact, and a forty year war on universities has left them without an institution capable of ruling on questions of theory. So instead there’s a sea of pundits unmoored from any touchpoints, making up reality as they go along.

      Similar things are happening on the left, particularly around issues of pro-corporate media. They’re just a decade or so behind.

      This isn’t to say that bias doesn’t exist: it does. But if you remove one institution that rules on truth and fiction you have to replace it with another. Extreme partisans have torn down the press and the universities, but they have not replaced it with a new apparatus for chasing down sources or doing experiments, or even the grunt work of careful reporting. They’ve replaced it with punditry, which now both creates realities and rules on their truth.

      I’m wondering if I’ve made this clear on this blog — I’m obsessed with fake news because it appears, in some ways, as the culmination of this decades-long process of tearing down the credibility of institutions — with an ample assist from the institutions themselves.

      • That’s a good point but not at all what I am getting at. Egypt used to have govt-endorsed “inaccurate” news; now it’s all over the place (i don’t think anything as extreme as completely made up, but real gross exaggeration and inaccuracy). The problem isn’t about authority of truth or knowledge. The problem is with human beings who aren’t prepared to deal with all of this. Which could be a problem of education, but not only traditional education. I don’t even know if my article explains what I mean, but it will be a series so I can keep clarifying

  2. Well, the point of the Digital Polarization Project is to try to help people deal with this (through education) so I’m sympathetic to the view. I must have an optimist that education can help us be up to the task under the cynicism.

    At the same time I think we can feel that quite a lot of this is being placed on people in a way that feels unfair and just bound to go wrong for a very long time.

    If I can use a not-quite-up-to-snuff analogy, I’d consider driving, which kills a lot of people but kills a lot less people than years ago. Part of that is we have learned to be better drivers. But part of it as well is that the companies building cars and the municipalities building roads were pushed to to see reducing accidents as part of their social responsibility. I’m not talking about *making* Facebook provide virtual seatbelts or network air bags but I think we need to think about what seatbelts look like in a social information system. And culturally we need to learn to be better drivers too — when I was a kid, we used to joke around about almost driving off the road while tired, but now with my own kid we treat tired driving in a way we used to reserve for drunk driving.

    The metaphor doesn’t fit because there are some complex skills involved as well as cultural norms. Sam Wineburg is going to have a post in Huffington Post (ironic I know) on this issue tomorrow that will be worth reading. I’m interested to read your piece as well. Let me know when it is out.

    • I am fascinated by thr driving/seatbelt analogy and I think it’s probably a really good one even though the complexity is different. It’s complex *enough* to be useful, w social elements as well as individual and moral ones. Egyptian women drive normally, but sometimes start late; patriarchal arguments against driving are “I trust you to drive well but don’t trust other people on the road”. Which is a fair argument, until you realize people who aren’t driving still trust someone else to drive FOR them, so that’s actually one more person you give control over to…so the argument breaks down and the solution would be to not go out at all (?). I went off on a tangent here but it might be a useful one. We don’t need to accept the premise of other people on the road, of course, or else we would all probably go crazy. Reading a master’s thesis (to examine soon) right now on cultivation theory which is also influencing me.

  3. Doing my best to follow this convo. I’m alarmed at the trouble we find ourselves in and appreciate the parsing & breaking down of real-time phenomena. Two ideas I’m taking with me: that we don’t experience fake news differently from.real news and what seatbelts might look like in a social information system. Compelling ideas I need to go think about. Thank you, both.

  4. Doing my best to follow this convo. I’m alarmed at the trouble we find ourselves in and appreciate the parsing & breaking down of real-time phenomena. Two ideas I’m taking with me: that we don’t experience fake news differently from.real news and what seatbelts might look like in a social information system. Compelling ideas I need to go think about. Thank you, both.

  5. Fake News is not the sole problem, but it does point to the need for teachers of all levels to be more immersed in teaching techniques to allow readers to make judgments on what they read (there is no real ‘truth’ out there, necessarily, since everything is written and read through filters — of our own making and of others).

    • I agree with you, and I think one very easy caveat to follow is that no one should put too much stock in reporting until it has been verified elsewhere. I have learned to stop posting anything that sounds too hyped or just seems ludicrous until I’ve had some time to delve into it or read someone who has done that already. No matter who it is, they can be wrong, so this works even with those pundits/media sites you already trust. Trust, but verify, I guess. hehe

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s