Epistemic status: My own theories that developed over years based on reading and giving mainstream media interviews about Quantified Self.

When the phrase 'fake news' became popular in 2016, many people called for critical news consumption. The American left used the term to talk about new venues that were very successful in producing stories that went viral on Facebook. Later Donald Trump co-opted the phrase to talk about the mainstream media.

The quality of the information of both kinds of sources is radically different. At the same time the mainstream media isn't a paragon of truth either. Given that there's little good writing about how to interpret news articles, I will try to describe a few key features of how reading the news media can mislead.

There's a common idea in the American discourse that you find informative articles by seeking for articles that are unbiased, have a neutral point of view and are objective. While the ideal has positive intent, following it leads to being less informed.

Why would you want to read an article from someone who doesn't have a neutral point of view? The average journalist has huge time pressure when he writes an article. An article that's based on the understanding of a topic that can be derived by researching the topic for a day isn't very deep because the author doesn't understand the topic very deeply. Traditionally, outlets such a Foreign Policy and The Economist, that take a point of view and are more informative than news media that pretends to be neutral. These articles are more thought through.  The author has an informed opinion on a matter – a bias that has intelligent reasoning behind it.

While some forms of bias can be easily seen by reading an article, many can not. It's easier to correct for bias when the author is open about his stance on the issue.

When Robert Moses took over the building of the Triborough Bridge there was a flaw in the previous plan. The sensible location for the Bridge would have been 100th Street as 85% of its predicted traffic was going to move through it. The plan for the Triborough Bridge called for it to be built at 125th Street which added 2,5 miles of additional travel for most users of the bridge. Robert Caro writes in The Power Broker in 1974, that William Randolph Hearst, who owned at the time two newspapers, wanted the bridge to be built on 125th Street, because he would financially benefit from it being built at that place as he owned deteriorating real estate that the government would buy from him as a result. 

Instead of picking a fight with Hearst, Moses allowed the bridge to be built where Hearst wanted it to be built. The project succeeded and the bridge opened in 1936. It would have been nearly impossible for a reader who tries to find out whether a related article in Hearst's newspapers is biased to find out about the conflict of interest, given that there wasn't any public information about it. This not only affects the reporting on the story of this particular newspaper but by doing what Hearst wanted in this instance, Moses could count on Hearst’s support for his other projects as well.

In contrast to a newspaper an organization like Amnesty International is far from having a neutral point of view. They invest a lot of effort into doing the research to get their facts right. This kind of research takes resources and the people that engage with it do it because they actually care about it and they have a point of view that it’s important. Even when you don't necessarily follow their conclusions they are often good sources for information.

The alternative to thinking of articles as either fulfilling the ideal of the detached and having neutral point of view is to practice an expanded version of theory of mind. Theory of mind is about thinking about what goes on the mind of someone else. Why "expanded" theory of mind? It’s about not only taking into account the mind of the author of the article. In most cases, the author of the article isn't the only person involved in writing the article.

In most newspapers both in print and online the headline isn't written by the person who wrote the article but by a headline writer. The headline is written by a person who didn't do any research on the topic and whose job it is to advertise the article in the newspaper to get people to read it. If there are claims made in the headline that the article doesn't make, those claims should generally be disregarded. Headlines are designed to entice the reader but may not accurately represent the content.

The first article that was written about in the context of Quantified Self was syndicated to multiple websites and some of those had new image captions that painted me in a negative light that the journalist who wrote the main article didn't write. When you read an article you should put special caution into what's written in image captions as there's a good chance they didn't get written by a person who researched the main article. 

While some websites publish articles without an editor being involved, in many cases an editor influences articles before publication. If a journalist makes a claim that creates legal liability for the newspaper when it's wrong, the editor will require deep research from the journalist to prove that the claim is right. On the other hand, the editor won't care about whether the journalist gets the age of a person inside the article right and as a result it's a lot more likely that an article will misstate the age of an interviewed person. 

Whenever you ask yourself whether a claim that's inside an article is well researched it's useful to think about how much the journalist and his newspaper has to lose when the claim will get shown to be wrong.

While journalists do provide supporting evidence for their claims to the reader, they frequently can't share all the evidence they have. During interviews there are three basic modes:

#1: Information can be attributed to the person who's interviewed

#2: Information can be cited by anonymizing the person "A trusted source within the administration said..."

#3: Information is given on deep background and can't be used unless the journalist finds another source for the information. 

A good investigative journalist usually has a lot of information that comes under #3 because people are more willing to share information with him under #3. 

John Carreyrou brought through his investigative reporting light on the problems at Theranos. One of the lead scientists at Theranos committed suicide and Carreyrou spoke to his wife. Given that the wife was under a non-disclosure-agreement she shared her knowledge with him under #3. Carreyrou had a lot more information about what Theranos did wrong than he could publicly disclose and the wife likely wasn't the only source of information that comes under #3.

Politicians who spent a lot of time with journalists often like it when they have someone to listen to talk about their troubles and when they don't want the information to appear in print it's often easy to simply declare the information to be shared under #3. There also cultural expectations that certain information should be automatically regarded as being shared under #3 in the mainstream media. When Michael Hastings wrote his article in the Rolling Stones that lead to the sacking of US general's Stanley McChrystal there was the sense that he violated the norms of the mainstream media who would have seen the information that got him fired has having been under #3 because of their cultural norms. 

Roughly, a decade ago Buzzfeed was founded as a new model of publishing. It's founder Jonah Peretti had the insight that readers not only come to a news website because they want to get the daily news but that the new social media technologies allowed readers to share articles with each other. Buzzfeed provided his writers metrics to tell them about how viral their stories were and started to incentivise writers by the number of clicks their articles received and how often their articles get shared. People share articles when they get angry or feel happy about an article. They don't share an article when it makes them sad. As a result the most photo series about the ruins of Detroit don't have people in them while the real Detroit is full of homeless people. If the reality is sad, a journalist will try to write a story that will make you angry rather than sad. 

Buzzfeed produced a model that allowed them to get a large audience but they missed respect. Buzzfeed decided that they want to be taken seriously and hired Michael Hastings and other investigative reporters. Even when the investigative reporters produce less clicks per hour of work, the stories they write are worth it because they give the website more prestige. That prestige in turn makes it easier for Buzzfeed to convince advertisers to pay for sponsored stories on Buzzfeed. It's fortunate that prestige often goes hand in hand with providing true information but in cases where it doesn't, the reader should be very careful about claims that are made.

Investigative reporting can sometimes get dangerous. Michael Hastings died at the age of 33 when he was according to his widow writing a profile of CIA Director John O. Brennan in questionable circumstances. It's not clear whether his death was an accident but whether or not it actually was an accident the incident likely makes journalists who go for similar stories nervous.

Trump's ties to the mafia are an interesting story that doesn't get much reporting. While there was enough evidence of such ties, Australian authorities blocked Trump from building a casino and the US media is generally very interested in writing negative stories about Trump, you find few stories about Trump's mafia connections. The problem with writing those stories is that it's not only an attack on Trump but also an attack on the members of the mafia who don't like to be attacked. Wikileaks hosts some interesting articles for anyone who wants to better understand how the related mechanism allegedly works. For obvious reasons, I won't recount specific facts in my article and thinking through my likely reasons is left as an exercise for the reader. 

The problem of missing information that a reporter can't access means that Forbes lists of the richest people is unlikely to actually list all the richest people, many of whom have a lot of their wealth in tax havens in complex structures where the ownership can't be assessed by a journalist. 

Even when the information is available, the real world is very complex. A journalist, who wants to reach a broad audience for his writing, has to keep his story as simple as possible. It's the mark of a great journalist to be able to take a complex issue and explain it in a way that's simple. Journalism doesn't work by the academic standards where claims get qualified to avoid saying anything that isn't strictly true. 

During my Quantified Self time, I gave a TV interview together with a friend who used the EmWave2 to meditate by doing heart rate variance biofeedback. We had the problem that most of the audience likely didn't know what heart rate variance was and there was no time to explain it because it wasn't central to the story of Quantified Self. After back and forth with the journalist to get the story to be more simple he ended up saying that he measures his heart rate. For the average member of the audience that was likely as informative as the more accurate claim that the EmWave2 measures heart rate variance but more sophisticated members of the audience would have had a problem if they would have interpreted the claim as face value. 

Thinking about complexity is important when thinking about public relations. When we discussed which questions to have in the LW census we thought about adding questions about illegal drugs but decided against it. Even a number that 14% of the LessWrong population used illegal drugs would constitute an interesting factoid for a journalist who writes a story about LessWrong even when the number equals the number of the general population. On the other hand, the fact that I argue against adding a question about measuring illegal drug use via our census is on it's own no factoid that's interesting to write about in a newspaper article. 

Savvy authors who write for a specialized audience about an important subject matter that's unlikely to be well received by a mainstream audience are well advised to raise the complexity of their writing in a way that they can't easily be quoted without the journalist explaining their position in more detail. While this doesn’t make it impossible to write a story, it makes it harder.

Seymour Hersh who won the 1970 Pulitzer Prize for International Reporting exposing the My Lai Massacre and its cover-up during the Vietnam War and who broke the story about Abu Ghraib wrote The Killing of Osama bin Laden. Hersh describes how the official story of how the US killed Osama bin-Laden is a huge fabrication. At the time when the US military in the official story killed bin-Laden, he was a prisoner of the Pakastani’s. The Pakastani’s killed him before the American helicopters arrived and flew his dead body away. Despite his track record as an investigative journalist Hersh couldn't get the article published in a normal mainstream venue and had to go to the London Review of Books to get it out. 

When we get back to stories for a mainstream audience, a story for a mainstream audience has to talk about narratives towards which the audience can relate. The question of how we deal with technology is culturally interesting and as a result journalists were very interested in writing articles about Quantified Self. It deals with our human relationship to the technology in our lives. 

When Egypt had its revolution our news media liked to talk about Twitter and Facebook because the idea that Twitter and Facebook change our lives is a narrative that Western readers can relate to. Western readers are less interested in the fact that the military who runs parts of Egypts economy had an interested in the revolution succeeding. 

They didn’t like that Hosni Mubarak and his son Gamal Mubarak where opening up Egypt for international business. While Hosni Mubarak served in the military, his son whom he groomed to be the next president had his first job at bank of America. Under him the military was going to lose power over the Egyptian economy to foreigners.

While the military could have decided to stop the revolution they didn’t need to because their power was never threatened. I wasn’t surprised when the military later reinstated the dictatorship because they never gave up any power. From the narratives of Western media that event was unforeseeable. This narrative where the people have the power to overthrow a government regardless of what the military wants to happen in turn leads to bad public policy.

According to Michael Cieply who worked at the New York Times, in its editorial process journalists are occasionally asked by editors to map a narrative a year in advance and then find the facts to write the articles according to the pre-maped narrative.

Hans Rosling asked the US public, the US media, EU public and EU media questions about how well global development is going. When he asked “What percentage of the world's one-year old Children are vaccinated against measles? 20%, 50% or 80%?” only 6% of EU media members answered the correct answer of 80% with 8% of the EU public getting it right. They did much worse than a chimpanzee who picks the answer randomly. He got similar results for other questions about global development. While the newspaper stories are mostly fact-based, the narrative that they tell about the global world makes people who read them less informed about what goes on in the world.

When you read the news, don’t see the article you are reading in isolation. Try to see it in the context that produced it to draw understanding of our world from it.

New to LessWrong?

New Comment
15 comments, sorted by Click to highlight new comments since: Today at 12:50 AM

When you read the news, don’t see the article you are reading in isolation. Try to see it in the context that produced it to draw understanding of our world from it.

It is the simpliest answer. Totally correct, but I hoped for more, when clicking title.

  1. It is not enough to read news media to get the context, allowing to judge reliability of news. I'd propose to read some books, to take a bigger picture, to grasp historical process which we are part of. To grasp some idea of metaphysics of human history. I couldn't explain what it is, because it is too complex and difficult to explain, it seems like my neural networks have learnt some generalizations from a very wide context, but how it often happens with neural networks no one could explain how they work. It gives ideas of what is possible and what is not, what is probable and what isn't.
  2. We could move on level lower and search article for signs of bias. If we knew bias, we could undo it. The good example of the most blatant signs s the journalistic art of Rita Skeeter in Rowling's Potter. Speaking of which Rowling failed to show how to work with such a journalism, her heroes either ignored Rita or (like Harry in the last book) felt pain from inability to separate lies from facts. It is possible to extract facts from Skeeter's articles, to throw away everything else, and to see a picture which is as real as it is possible. But Potter dived into a reality pictured by Skeeter and then painfully tried to resurface. It was a mistake from the beginning (though understandable, Potter was 17 at that moment, and one of his hiddens powers to fight Voldemort was his emotionality).
  3. To measure bias we could evaluate not just the single article, but a lot of articles of the author of the article, or from the same publisher. To do it I systematically read news from several resources. Even when I cannot find nothing interesting for me, I pick some article at random and read it. This way you learn a style, you learn a narrative, you learn to predict what this media would write for any occasion, and your mispredictions will be very informative.
  4. Read different sources. Truly different. If you read NYT, add Fox News to your reading list, but do not stop there, add something completely different, something outside of the system you watching. It is very hard to understand a system being a part of it. Some believe that it is even impossible. Though it is a deep philosophy.
  5. Read a textbook on experimental psychology. I believe no science know about experiments as much as psychology. Physicists do not know as much, because they deal with simple systems without brains. They are not forced sometimes to abandon experiment and use pseudo-experimental plans, or other reseach plans. They need not to invent a new physics each few years, so they know how to exist with a ready-made theory, but they know a little how to deal with reality where all theories were failed. Sociologists deal with too complex phenomena to be successful, and they mostly stuck with correlational research plans (so no reasoning about causality). Psychology is in the between, it could struggle with inability to devise an experimental plan, and at the same time it have a chance to devise it and to test it on real subjects. Textbook on experimental psychology could give you the theory and some practice to learn how to evaluate a validity of a research. It helps a lot even if I read an anonymous comment in internet, not just when I'm reading a psychological paper.

We could move on level lower and search article for signs of bias. If we knew bias, we could undo it.

It's easy to know that the New York Times and Fox news both have a pro-corporate bias. If you however are not exposed to the knowledge of how the Egyptian military competes against corporate actors in their economy you can't undo the bias when reading either source on the Egyptian revolution. 

If you read NYT, add Fox News to your reading list, but do not stop there, add something completely different, something outside of the system you watching. 

I'm skeptical that orienting yourself towards big news publishers is a good. One of the main problems is that their articles intend to stand on their own and usually don't link out much. A lot of articles contain little important information.

I either read articles because someone recommends an article or because I research a given topic and then go to Wikipedia/Google/StackExchange.

Read a textbook on experimental psychology. I believe no science know about experiments as much as psychology. 

Do you have one that you would recommend in this context? I consider Stanovich's How to Think Straight About Psychology good in explaining how knowledge can be aquired in psychology. Gwern's essay on Causation is worth reading to remind oneselves that knowing things about the world is really hard and it's easy to be mislead. 

My own answer is to make predictions (commonly on Good Judgment Open), and then notice a couple of months afterwards that the events which the news predicted (Mark Esper or Paul Guedes quitting or getting fired, Lukashenko being overthrown, etc.) tend not to happen.

Hmm... I think the main problem with fake news is mostly that it is propagated and spread/forwarded by exactly those people who don't think they need to do any research coz they already "know" the "truth" anyway. There's a whole segment of society for whom the word nuance is a waste of time. We all know that becoming more informed and/or finding out another point of view can be easily accomplished merely at the cost of some time and attention; unfortunately, a sizeable part of the public does not feel the need to do so.

This is mostly because the dedicated pursuit of the Truth is not the objective (or the reward) of consuming content online - for many, I suspect it is more to support the narrative they have to explain the world around them. More often than not, such a narrative is only partially true, and only sporadically adjusted or updated with new and/or hitherto unknown facts.

For example, let us suppose you have a populist point of view on Muslims, and tend to "assume" as true claims like them tacitly supporting terrorist attacks, treating their women in medieval ways, and perhaps being very in-group oriented, so not mixing much with the locals.

Then one day a Pakistani family moves into your street right next door, and after a few months you can't help but realize that, actually, these are really nice people, always courteous, the mother is always cheerful, and you overheard the father talk to the mailman both expressing what sounded like sincere disgust about the latest terrorist attack in France.

What is likely to happen next?
1) You adjust your opinion only about the family; your attitude to the rest of Muslims remains the same;
2) You change your mind about Pakistani Muslims; your attitude to the rest of Muslims remains the same;
3) You realize you're an ignorant xenophobe and go online to properly educate yourself on Islam culture;
4) After eating some humble-pie, you approach the family, confess you're embarrassed for being so ignorant and ask them to help you understand how Islam and/or Pakistani culture works.

Ok so you are not likely to do the last one... but the best response, surely, is nr. 3. Yet how many people will choose that? I "know" plenty of people that have a disparaging or... well - "out-group causing" opinions of them forrinners, but do tend to say they know a few of "them" who are OK. Apparently, without any coherent explanation as to how (or why) the "good ones" differ from the "bad ones".

Also, consider peer pressure. Choosing answer (1) allows the rest of your world to stay "as is"; your nephew can still forward you the insulting Muslim memes, you can still get outraged in the same way each time some Islamist attack occurs, etc. - the influence of habit and social circle also plays a major role here... it kind of goads you into staying in your particular bubble of prejudice. Is ending your relationship with that otherwise very close nephew worth being a bit less ignorant towards a group of mostly outsiders with whom you rarely interact...? For many, it wouldn't be...

If people valued nuance and genuine truth and facts, they know very well you can find them. Start on Wikipedia for all I care; and take it from there. Escape your algoritm-spun bubble of conspiracy loonies on YouTube you must. Intentionally seek the opinion of those in disagreement, you should. Take some time for this, however, they shan't. People are sceptical first and foremost towards information that indicates they might be wrong.

We all know that becoming more informed and/or finding out another point of view can be easily accomplished merely at the cost of some time and attention; unfortunately, a sizeable part of the public does not feel the need to do so.

Have you read the article? Are you saying it was easy for you to know that know that the Egypt military allowed the revolution to happen because they didn't like Mubarak and the other examples in the article?

Well - I think that you are obviously correct in stating what you state. However, the issue is not that people who read the news don't know about this; rather, that the group of people that is careless in evaluating the news they read can't be bothered to spend so much time corroborating, cross-referencing and classifying the stuff they read. Another way of saying what I mean is - here you're preaching to the converted who, I suspect, to a significant degree already apply prudence to the news they read.

I am operating on the assumption here that you wrote this text with the ultimate aim to improve the way in which people process and interpret the news, and I imagine that ideally you would want the proposed approach to limit the damage that reflexive emotional reactions to half-baked "news" can do. Unfortunately, the vast majority of those who constitute the bulk of the problem (the crowd that forwards nonsense and biased stuff, the people who act on it by forming stereotypes  etc.) - they are not exactly likely to adopt the approach you (rightfully) advocate; if they would be likely to do so (and so show particular concern for the truth) we would not have the problem of fake news in the first place, at least not at the current scale.

When I write a post on LessWrong I'm writing for an audience that's already engaging in rational thought. Knowing things is hard the fact that you suggests it's easy is to me an indication that suggests you don't really get the point. 

What's exactly "the current scale" of it? In the 40s and 50s Robert Moses happened to be the most powerful person in New York politics without it being able to be understood how that power was wielded by reading any of the newspapers.

Are there people today who have comparable power to Moses back then but don't appear in any newspaper? It's intrinsicly a hard question to answer. Maybe we can observe that nobody manage to get as much done as Moses got done, so power can't be as concentrated, but thinking about whether that's good or bad isn't as straightforward.

I guess what I meant by “easy” is compared to not doing any fact checking. So, 2-5 minutes of additional searching/additional sources would often be sufficient to realize something is most likely biased and/or fake news in, say, 80% of cases. It’s quite sad and discouraging that so many readers are unwilling to do even that, though again if the aim is confirmation of world view and not highest probability of accuracy and truth... it actually “makes sense” to not check ;)

It’s quite sad and discouraging that so many readers are unwilling to do even that

I think it's eminently reasonable to avoid spending time on learning facts about the world that come to your attention as popular news, and not because you are interested in them. Epistemic hygiene needed to render the possibly biased and confused news articles harmless shouldn't in itself require additional research, however minimal. Otherwise it won't work in situations where that research is not immediately feasible, and then you are truly in peril.

I had to read that twice to make sure I figured out what your point is :) Alright - well, look. The truth is, of course, that 95% of the news you read is utterly and completely irrelevant to you in any impactful sense. Try not following news for a month - you will soon realize you have not actually missed anything. Well - now with COVID this might be a little different, but only a little.


If you do follow news, it would be proper and prudent to at least care about the veracity of it, especially if you have a habit of forming opinions about said news, and, in particular, if you somehow do find the time and energy to spread this opinion online. The combination: "Follow all sorts of news" + "Form ignorant opinions" + "Spread my ignorant opinions" is not a good one. Worse things of course have happened ;) but that is certainly not an admirable state to be in.

And if people read news to, you know, feel like they are aware of what's going on in the world, however superficially, you would want to expend at least a minimal/nominal effort to ensure at least some confirmation of the veracity of what you are reading. For example, to have a moderately sensible opinion about Black Lives Matter, you don't need to study the entire US history... but reading a serious article or two (ideally from different sources) would be appropriate to make sure you at least know more than one point of view (especially since your point of view is probably "pre-confirmationbiased" by your Google bubble ;)

but reading a serious article or two (ideally from different sources) would be appropriate to make sure you at least know more than one point of view (especially since your point of view is probably "pre-confirmationbiased" by your Google bubble ;)

"Knowing more then one point of view" is a goal that's distinct from knowing the truth. 

For example, to have a moderately sensible opinion about Black Lives Matter

It seems to me that there are a lot of different issues involved in the hashtag #BlackLivesMatter and if you try to understand what's happening through the complexity of a hashtag.

One of the interesting things about it is that when the key BlackLivesMatter organization has an about page about what their organization is about. It spents more time critizing traditional Black Liberation movements then directly critizing police brutality. 

Power changes in Black politics is however not a topic that anybody has a real interest to tell us about. Maybe, there will be a good book in a few years that recounts our contemporary history but at the moment the information that's publically available seems  very superficial. 

We can observe about pages and attacks on King's legacy by bringing up old tapes, but those are only the things that stick out and there's a lot of room that's unknown.

When it comes to understanding hashtag politics it's worth noting that we are one year after #MeToo at the point where two US candidates with multiple credible sexual assault charges are fighting each other. It's embarrasing enough that it's not often talked. It's like the Skulls&Bones vs. Skulls&Bones election of 2004 which would be a great narrative but nobody is interested to talk in those terms about this election. 

Oh I don't really "do" Twitter actually... nor Facebook since about a year. Now and again one of my friends shares and tweet and sometimes it can be an interesting start of a topic but... though I've been doing Internet since 1995, Twitter is just too vacuous for my liking. In response, now and again I'll send a 1 hour+ YouTube link back ;)

And yes of course, multiple points of view need not bring one close to the Truth, however...

In a large number of narratives, especially, it seems, the most relevant ones, finding the truth may be practically impossible, and sometimes there simply is no truth, or at least not just one. To some people aspect X is irrelevant, others might believe it crucial. This news network claims Witness Y is credible, some other one calls him a corporate shill. Unless you would be able to get into the minds of each human involved, what you end up believing is the truth will always be an approximation.

Take for instance the recently more often occurring phenomenon of "influencers" (shudder) bloggers or journalists looking into the obscure past of what someone who is having his/her 15 minutes of fame has posted back in, say, 2004 on some now-defunct blog, and bleating out on Twitter anything remotely controversial or tentatively indicative of hypocrisy. I doubt you will ever settle the debate whether people can genuinely change or not. I know that I've had views I no longer hold today - both "benign" and "tough love" ones... and while previously held view will always have the familiarity bias, they can actually be genuinely a thing of the past. Yet if they are found online and are at odds with what I would be saying today, poof there goes half of my credibility...

And - getting multiple points of view at the very least will give you some idea why certain people apparently seem to find a given topic or story important. The net outcome may well be that you will be further from the truth, swimming in a sea of conflicting interests... and yet, still understand the nature of the issue in more detail :)

If you do follow news, it would be proper and prudent to at least care about the veracity of it

My point is that there is an important skill that should allow you to passively observe unreliable stories without being harmed by them. It's possible for this skill to give you more than nothing, so its application to news is not equivalent to not following news. In particular, you can notice something potentially interesting or useful and research it, but there should be no need to do any research. If there is such a need, it should be satisfied by improving generally applicable epistemic defenses, not by fighting off specific news articles.

Yeah alright... I guess you could call that passive casual observance :)

That's the traditional option. A simple improvement is to compartmentalize ideas (according to origin and epistemic status) while still taking them seriously within each cluster (thinking about their implications in a lawful gears-level way). I'm guessing this is how Judaism works.