The internet and confirmation bias

The issue of confirmation bias has come up repeatedly here before. Basically, people evaluate new information in a way that is far from impartial; new information that seems to confirm pre-existing beliefs is generally filed as evidence for the appropriateness of those beliefs, while contradictory information is downplayed or ignored. While this phenomenon is ancient, there does seem to be good reason to think that it may be especially acute now, as the media becomes more personalized and segmented.

That danger is highlighted by Harvard academic Ethan Zuckerman, who gave a TED talk on how social networks mislead us. Because we are exposed to the thoughts of people who are already much like us, we are at risk of being convinced that we are more typically than we really are, and our views are more mainstream and justified than may actually be the case.

How much of a problem would people say this is, both from the perspective of being well-informed citizens and in the context of being effective in promoting particular policies? Is there any way either social networks or individuals can combat this entrenching of confirmation bias? For my own sake, I have been trying to incorporate more articles from newspapers I disagree with into my daily reading.

Author: Milan

In the spring of 2005, I graduated from the University of British Columbia with a degree in International Relations and a general focus in the area of environmental politics. In the fall of 2005, I began reading for an M.Phil in IR at Wadham College, Oxford. Outside school, I am very interested in photography, writing, and the outdoors. I am writing this blog to keep in touch with friends and family around the world, provide a more personal view of graduate student life in Oxford, and pass on some lessons I've learned here.

20 thoughts on “The internet and confirmation bias”

  1. I am generally skeptical of these claims about the “internet echo chamber”, since I don’t think we are actually able to filter the internet that perfectly, and these claims are rarely based on evidence. For example, research by the Pew Internet and American Life Project shows that those who spend the most time online are the most exposed to opposing views.

  2. The idea of confirmation bias is a conspiracy of the left-wing media to discredit real news sources like Fox. If only more people watched Fox you would understand this too.

  3. That talk also points out the role of mainstream media in creating a local bubble – the reduction of international news from 35% of broadcasts in the 70s to 12% today.

    There is something a bit offensive about the way he talks about flight patterns, flights, and the lack of substantive connectivity between South America and Africa, without pointing out that flying is basically reserved for the world’s super-rich.

    Why not, instead of reading a Canadian or American newspaper you disagree with, read a Newspaper from somewhere else in the world?

  4. It seems deeply problematic to me that someone would be so concerned with confirmation bias, and yet say any movement is a failure if it is not immediately well received by the mainstream media.

    What is the difference between what is here called “confirmation bias”, and what is elsewhere called “the manufacture of consent”? I.e. why is a claim like this:

    “It’s one of the great assumptions underlying modern democracy that an informed citizenry is preferable to an uninformed one. “Whenever the people are well-informed, they can be trusted with their own government,” Thomas Jefferson wrote in 1789. This notion, carried down through the years, underlies everything from humble political pamphlets to presidential debates to the very notion of a free press.”
    (from the Boston Globe story linked above)

    Not immediately linked to

    “That the manufacture of consent is capable of great refinements no one, I think, denies. The process by which public opinions arise is certainly no less intricate than it has appeared in these pages, and the opportunities for manipulation open to anyone who understands the process are plain enough. . . . [a]s a result of psychological research, coupled with the modern means of communication, the practice of democracy has turned a corner. A revolution is taking place, infinitely more significant than any shifting of economic power. . . . Under the impact of propaganda, not necessarily in the sinister meaning of the word alone, the old constants of our thinking have become variables. It is no longer possible, for example, to believe in the original dogma of democracy; that the knowledge needed for the management of human affairs comes up spontaneously from the human heart. Where we act on that theory we expose ourselves to self-deception, and to forms of persuasion that we cannot verify. It has been demonstrated that we cannot rely upon intuition, conscience, or the accidents of casual opinion if we are to deal with the world beyond our reach.
    – Walter Lippmann, Public Opinion”

    If you’re worried about confirmation bias – you should be worried about the way confirmation bias is used by powerful institutions to make sure we for the most part believe the lies our governments tell us.

    It’s not strange in light of this, that a common emotion in the counter culture is “Rage”. We can debate about whether it’s positive or negative, or effective or alienating. But we can’t pretend it to be odd that given the incredible ability of the system to prevent dissenting opinions from gaining any public credence, that a predictable response would be to smash things.

  5. I think there is a serious problem with the criticism levelled in the TED talk referenced above.

    Sure, through the internet you can choose who to connect with, and form groups cross province, even cross country, who care about similar issues and can then mobilize in favour of them.

    This is a bad thing? I just found out about a group of Quebec nationalists who are calling for an inquiry into the G20. I would not have so quickly found out about this if it were not for my facebook feed. But, there is lots I didn’t see.

    There is always lots you don’t see. Everything you see is not seeing a million other things. And sure, you shouldn’t get stuck in a rut, as I probably am on this whole “giving a shit about justice” tirade.

    That said, the “echo chamber” complaint about the internet is at least something people can choose to get out of, by spending two seconds on google and looking up foreign newspapers. The echo chamber effect of the popular media, i.e. pre-internet, is much more difficult to escape.

  6. Here is an example of people moving beyond their usual sources of information. Well-known deniers WUWT and Jo Nova both posted pieces based on a media release from BNP (far right extremist/racist hate party in the UK), before later pulling them when they realised that not all sources are worth exercising your confirmation bias on.

    The cached entries are here and here. Unsurprisingly, there were no other sources for this claim of evil eco-totalitarianism outside of the BNP.

  7. Confirmation bias does seem to be a problem with splintered and personalized media. The public discussion about policy can get so fragmented that people can no longer really comprehend one another, and become caricatures to each other as a consequence.

  8. ” A generation of digital activists had hoped that the web would connect groups separated in the real world. The internet was supposed to transcend colour, social identity and national borders. But research suggests that the internet is not so radical. People are online what they are offline: divided, and slow to build bridges.

    This summer Ms Boyd heard from a scholar in Brazil who, after reading her research, saw a parallel. Almost 80% of internet users in Brazil use Orkut, a social network owned by Google. As internet use rises in Brazil and reaches new social groups, better-off Brazilians are leaving Orkut for Facebook. That is partly because they have more friends abroad (with whom they link via Facebook) and partly snobbishness. Posh Brazilians have a new word: orkutificação, or becoming “orkutised”. A place undergoing orkutificação is full of strangers, open to anyone. Brazilians are now the second biggest users of the micro-blogging site Twitter; but some wonder whether the dreaded o-word awaits that neighbourhood too.”

  9. Pingback: The Young Turks
  10. “Both anger and contempt have deep psychological roots. Anger usually stems from feelings of unfairness or betrayal. Contempt is anger mixed with disgust. Anger and contempt are not just emotions. They are scripts that determine our political conversation. If you are a conservative blogger, you will hunt out material that shows liberals to be unpatriotic and dangerous, because your audience wants affirmation of its underlying feelings. If you are liberal, you will play up material that shows conservatives to be stupid, because your audience wants affirmation of its sense of superiority.

    This selective weeding does not always happen consciously. Just as interpersonal relationships have unconscious scripts that determine how people talk and argue, so, too, our national conversation follows scripts that are more powerful because they are unconscious. When we treat one another badly because we mean to behave badly, we know we are doing something wrong. When scripts inside our hidden brains make us treat others badly, we do not realize we are behaving badly. Couples in distress who come to marriage counselors are invariably blind to how their internal scripts are helping to produce the problems that mark their daily lives. If conservative America and liberal America were married to one another, this relationship would have one side wearing overalls, swilling beer, and screaming all the time, while the other sipped cappuccinos, read The New Yorker, and pretended it couldn’t hear a word. “

  11. “A host of psychological experiments demonstrates that it doesn’t work like this. Instead of performing a rational cost-benefit analysis, we accept information which confirms our identity and values, and reject information that conflicts with them. We mould our thinking around our social identity, protecting it from serious challenge. Confronting people with inconvenient facts is likely only to harden their resistance to change.

    Our social identity is shaped by values which psychologists classify as either extrinsic or intrinsic. Extrinsic values concern status and self-advancement. People with a strong set of extrinsic values fixate on how others see them. They cherish financial success, image and fame. Intrinsic values concern relationships with friends, family and community, and self-acceptance. Those who have a strong set of intrinsic values are not dependent on praise or rewards from other people. They have beliefs which transcend their self-interest. “

  12. So what happened? In a phrase, “confirmation bias.”

    All people tend to embrace without question information that squares with existing beliefs, but the strength of this bias varies according to the strength of the existing belief.

    Thus, if you are a fiercely committed Conservative and you see a shot at the Liberal leader, that’s good enough. It doesn’t occur to you to ask if it makes sense, much less to Google the source. And what sort of person is more likely to be a fiercely committed partisan? Not the stupid and ignorant. They’re more likely to be apathetic. It’s the smart and informed.

    Fine, you may think. Smart and informed partisans may not question information that supports their beliefs. But what happens when contrary information is put under their noses? Surely then they will revise their opinions accordingly.

    Well, we can test that now. After reading this column, will any fiercely committed Conservative partisans read the Maclean’s article, protest their party’s misleading ad, and change their opinion of Michael Ignatieff?

    No. They will rationalize: They will stitch together an explanation that allows them to see the Conservative ad as less misleading than it is and Ignatieff’s comments as less simpatico than they are. Once they have that, their cognitive dissonance will melt like snow in the sun.

    Everyone can cobble together self-serving explanations but some are better at it than others. The superior rationalizers? Again, it’s not the stupid and ignorant. It’s the smart and informed.

  13. Effective leaders such as David Petraeus understand decentralization, experimentation, and adaptation. And they understand that giving orders from the top of a pyramid is the wrong way to lead. “Eric Schmidt, Google’s chairman and, until recently, its CEO, views his role at Google as mediating a debate and forcing other people to reach decisions, rather than making decisions himself,” notes Harford.

    Now compare that to the government of Canada. Successive prime ministers have all but erased parliamentary governance. In its place is a pyramid, at the top of which is the Prime Minister’s Office. Enthroned in the PMO, the PM is a pharaoh.

    Blame the PMs. But blame also ourselves. In our political culture, a leader who acknowledges uncertainty and encourages experiments is “indecisive.” A leader who permits dissent is “weak.” A leader who changes his mind in response to new evidence is a “flip-flopper.”

    A real leader is one who centralizes power, is certain of everything, who breaks the knuckles of anyone who disagrees, who never admits to being wrong, and who will deny to his last breath ever having changed his mind about anything. A real leader is a Great Man issuing orders from the top of a pyramid.

  14. The Filter Bubble: What the Internet is Hiding From You. By Eli Pariser. Penguin Press; 294 pages; $25.95. Viking; £12.99. Buy from,

    ELI PARISER is worried. Why? Call a friend in another city or a foreign country, and ask them to Google something at the same time as you. The results will be different, because Google takes your location, your past searches and many other factors into account when you type in a query. In other words, it personalises the results. As Larry Page, the chief executive of Google, once put it, “the ultimate search engine would understand exactly what you mean, and give back exactly what you want.” Eric Schmidt, executive chairman of Google, muses that someday it might be possible for people to ask Google which college they should apply for, or which book to read next.

    This is only one example of internet personalisation. Mr Pariser, an internet activist best known as a leading light at, a progressive online campaign group, sees this as a dangerous development. Netflix, Amazon and Pandora can predict with astonishing accuracy whether you will enjoy a particular film, book or album, and make appropriate recommendations. Facebook shows you updates from the friends you interact with the most, filtering out people with whom you have less in common. “My sense of unease crystallised when I noticed that my conservative friends had disappeared from my Facebook page,” Mr Pariser writes. The result is a “filter bubble”, which he defines as “a unique universe of information for each of us”, meaning that we are less likely to encounter information online that challenges our existing views or sparks serendipitous connections. “A world constructed from the familiar is a world in which there’s nothing to learn,” Mr Pariser declares. He calls this “invisible autopropaganda, indoctrinating us with our own ideas”.

Leave a Reply

Your email address will not be published. Required fields are marked *