Related:Bounded Distrust, Moderation Is Different From Censorship

I.

With a title like that, obviously I will be making a nitpicky technical point. I’ll start by making the point, then explain why I think it matters.

The point is: the media rarely lies explicitly and directly. Reporters rarely say specific things they know to be false. When the media misinforms people, it does so by misinterpreting things, excluding context, or signal-boosting some events while ignoring others, not by participating in some bright-line category called “misinformation”.

Let me give a few examples from both the alternative and establishment medias.

II.

Looking on Infowars, one of their top stories as I write this is New Vaccine Data Shows Alarming Number Of Stillbirths And Miscarriages Caused By Covid Shot. It’s based on the VAERS system of vaccine adverse events, which looks like this:

Twitter avatar for @Storm4CongressJason “Storm Chaser” Nelson (Shadow-Banned) @Storm4CongressVAERS data since 1990: Stillbirths & miscarriages as attributed to a vaccine reaction Image[5:03 AM ∙ Oct 25, 2022

7,332Likes4,179Retweets](https://twitter.com/Storm4Congress/status/1584772554048229376)

They draw the predictable conclusion that COVID vaccines caused the spike in 2021 and 2022 and therefore are very dangerous. I haven’t looked into this deeply, but I’m pretty sure I can guess what happened - you’re supposed to report adverse events (like stillbirths) to VAERS if they happened within some amount of time of a patient getting a vaccine. In 2021, probably more pregnant women got vaccines than in previous years [edited in response tohere], and they were much more nervous about them and more aware of the reporting system. If some constant number of pregnant women have stillbirths, but women and doctors were more willing to link it to the scary new vaccination, then the number of adverse events (eg stillbirths within X time after getting vaccinated) should go up a lot in 2021, which it looks like it did.

Or how about this article: Kari Lake Trial Bombshell: Audit Reveals 42.5% of Ballots Randomly Sampled Were ILLEGAL Ballots. This describes the results of a real audit. It looks like 42.5% of ballots audited were printed the wrong size (19 inch ballots on 20 inch paper); if there’s some regulation saying the ballots have to be the right size, I guess this might be “ILLEGAL”. The Lake campaign says this proves some kind of foul play. The government argues that maybe they messed up printing the ballots, but it was an honest mistake, and if it rendered the ballots machine-unreadable they would have counted them by hand later, so the results are still fine. Infowars does a terrible job providing balance; they mostly just quote the Lake campaign’s accusations that it was sinister, while ignoring or downplaying the more plausible government statements. Still, on the most nitpicky level, as far as I can tell the article doesn’t say a lot which is literally false. Perhaps great journalism would investigate how the printing process worked and where it went wrong, but as far as I know neither side does that - they just report the relevant officials’ claims in more vs. less accusatory tones and expect you to make a judgment call based on your priors (mine are on “honest mistake”).

Looking through Infowars, it looks like many of their articles are around this quality of these two. Others are even less misinformative; there are lots of articles about members of some group Infowars doesn’t like committing a crime or doing an offensive thing; usually other sources confirm that these crimes or offenses are real. Or articles about “EXPERT SAYS X!”, where someone who could be charitably described as an expert (an MD or PhD) really did say X, even though X is insane and all other experts disagree. If Infowars is lying here, it’s by choosing to report on these stories instead of others, in a way that suggests they’re important - not by making up completely imaginary things on the spot.

(if you disagree with this, it might be worth looking through the front page of www.infowars.com and calculating what percent of the articles seem technically-true-but-misleading vs. completely-made-up. I tried this and had trouble finding the latter, but your experience might differ)

III.

So Infowars often provides accurate data, but interprets it incorrectly, without necessary context. They’re not alone in this; it’s much like how the New York Times reports on real child EEG data but interprets it incorrectly, or how Scientific American reports real data on women in STEM but interprets it incorrectly, etc. This doesn’t mean these establishment papers are exactly as bad as Infowars; just that when they do err, it’s by committing a more venial version of the same sin Infowars commits.

I criticized this story - repeated by Mic, New Republic, and the Washington Post - saying that only 0.01% of welfare users tested positive for drugs. If true, welfare recipients would use drugs at less than 1% of the rate of the general population - and, the articles heavily implied - conservatives worried about people spending their welfare money on drugs were therefore unscientific and bigoted. None of the stories mentioned that the “test” was just asking the welfare recipients whether they were taking drugs, with the threat of taking their welfare away if they said yes, and no attempt to check whether or not they were lying. A few of the articles mentioned a different attempt at urine drug tests, which only a few recipients failed - but didn’t mention that they had the option of not taking the drug tests and that many people (probably including all the drug users) chose not to take it. Some would say this is important context! But again, there are no outright lies - 0.01% was the true result of this (very stupid) test.

Or consider this New York Times article (which I’ve also criticized before): Free Market For Education: Economists Don’t Buy It. It said that only 36% of economists on a survey supported school vouchers - and if even economists don’t support a free market policy, surely that policy must be very stupid indeed. Not mentioned in the article: only 19% of economists in the same survey opposed school vouchers. The majority described themselves as uncertain - but among those who expressed an opinion, nearly twice as many were pro as con. Again, some might say this was important context. But NYT didn’t lie outright; they reported the headline number correctly.

IV.

Okay, that’s my nitpicky point. Who cares? Obviously all of this kind of stuff is more than deceptive enough to in fact leave a bunch of people misinformed. So why do I care if it misinforms them by lying, or by misinterpreting things and taking them out of context?

I care because there’s a lazy argument for censorship which goes: don’t worry, we’re not going to censor honest disagreement. We just want to do you a favor by getting rid of misinformation, liars saying completely false things. Once everybody has been given the true facts - which we can do in a totally objective, unbiased way - then we can freely debate how to interpret those facts.

But people - including the very worst perpetrators of misinformation - very rarely say false facts. Instead, they say true things without enough context. But nobody will ever agree what context is necessary and which context is redundant.

For example, suppose Infowars claimed that police shootings in the US cannot be racially motivated, because police shoot slightly more white people each year than black people (this is true). This is missing important context: there are ~5x as many white people in the US as black people, so police shooting only slightly more white people suggests that police are shooting black people at ~5x higher rates. But I claim it’s also a failure of contextualization when NYT claims police shootings must be racially motivated because they happen to black people at a 5x higher rate, without adding the context that police are called to black neighborhoods at about a 5x higher rate and so have no more likelihood per encounter of shooting a black person than a white person. Perhaps the failure to add context is an honest mistake, perhaps a devious plot to manipulate the populace - but the two cases stand or fall together with each other, and with other failures of contextualization like Infowars’ vaccine adverse response data.

But lots of people seem to think that Infowars deserves to be censored for asserting lots of things like their context-sparse vaccine data claim, but NYT doesn’t deserve to be censored for asserting lots of things like their context-sparse police shooting claim. I don’t see a huge difference in the level of deceptiveness here. Maybe you disagree and do think that one is worse than the other. But I would argue this is honest disagreement - exactly the sort of disagreement that needs to be resolved by the marketplace of ideas, rather than by there being some easy objective definition of “enough context” which a censor can interpret mechanically in some fair, value-neutral way.

Nobody will ever be able to provide 100% of relevant context for any story. It’s an editorial decision which caveats to include and how many possible objections to address. But that means there isn’t a bright-line distinction between “misinformation” (stories that don’t include enough context) and “good information” (stories that do include enough context). Censorship - even the “safe” kind of censorship that just blocks “fake news” - will always involve a judgment call by a person in power enforcing their values.