Confirmation bias: 6 ways to recognise it and 5 ways to counter it

confirmation bias

Much has been made of ‘confirmation bias’, but much less has been made of what it is, how it works, how to recognise it in yourself and how to counter it. It is a concept which has been pushed hard in the political discourse over the last few years, ladled on situations where one side perceives the other side as being unreasonable. And, if one must admit to it, it’s done so in a very token way, merely paid lip service before continuing unabated by the said concept itself. It’s become ever more relevant as social media algorithms create echo chambers of information and the objective pursuit of facts and truths becomes ever more fraught.

It’s important to be under no illusions of how corruptive cognitive biases are, as was clearly articulated in this Psychology Today article – which is highly quotable on the topic – and describes confirmation bias as:

“The direct influence of desire on beliefs… motivated by wishful thinking… we embrace information that confirms that view while ignoring, or rejecting, information that casts doubt on it… we may become prisoners of our assumptions… wishful thinking is a form of self-deception… self-deception can be like a drug, numbing you from harsh reality, or turning a blind eye to the tough matter for gathering evidence and thinking… it feels strong and counterintuitive to look for evidence that contradicts our beliefs. This explains why opinions survive and spread. Disconfirming instances are far more powerful in establishing the truth… a true definition of self-confidence: the ability to look at the world without the need to look for instances that pleases your ego.”

It’s no accident that the author of that article is a professor of addiction.

It’s also a topic which gets right to the heart of free speech and national discourse, a democratic society functions on the ability to conduct a effective countrywide discussion from which a decision can follow, and much of what is now referred to as “deep divisions” or “a democratic crisis” seems to come down to this central ability slipping away from us.

To those of us who studied a science degree, the idea of cognitive biases is not new. There are a number of them, including ‘fundamental attribution error’ – explaining a certain observation as the cause of an underlying phenomena which is actually not responsible for the observation; self-serving bias – tendency to claim more responsibility for successes than failures; belief bias – mistaking the fortitude of your belief for its logical strength; framing – taking too narrow a view of the relevant context; and hindsight bias – the “I-knew-it-all-along” effect. Confirmation bias is also referred to, perhaps preferably, as ‘myside bias’.

There are many subjects on which science can’t respond, and are the purview of history or art, for example – cognitive biases, however, are not only the realm of science but something scientists learn to identify and contend with by profession. An ability to do this would surely be unavoidable in any definition of a scientist or science. The recognition of how powerful these biases are is particularly acute when studying psychology, firstly because it’s the study of the very organ which has endowed you with the inadequacies to achieve objective understanding. And secondly because the room for subjectivity is much larger when studying abstract and indefinable notions like “self-esteem”, a point which is also common to business and political notions, hence it’s relevance here.

The whole phenomena has taken an implicitly ugly tone in much political discourse, which this article in The Guardian is a good example of; it focuses hard on illustrating the idea of confirmation bias, and would never be so arrogant as to suggest that its own leftist views are uncontaminated by it, but conveniently seems to use a herd of right-wing-favoured fallacies as part of its illustration. Political commentators and pundits currently have a tendency to acknowledge their cognitive biases before moving swiftly on – having not dealt with them – to describe how prevalent and pernicious those same biases are in their opposition. Some of the worlds most prestigious publications are regularly guilty of this.

It’s important to engage fully with the concept and recognise its potency: disabuse yourself of the notion that you are immune from or in any way above this phenomena, or that you can, without effort, recognise when it’s unconscious activities are at play. It matters not how detailed your understanding is; it matters how you approach the information. All of this is not to exclude the importance of personal stories and experiences, but they should be recognised as such – just as objective, ubiquitous truths should not replace experiential knowledge. Both are important provided they are dealt with appropriately. But the human brain is built for stories, and not for objectivity.

Recognitions of cognitive biases

There are a number of symptoms of cognitive biases which manifest themselves in various ways, all of which aren’t difficult to guard against but will always be the path down which one’s mind most naturally travels. Recognising when it’s happening is often all that’s required to enact a course correction.

They never kick in

Cognitive biases are not something which, on the whole, have an introductory moment. There’s no feeling or sense you can look out for or identify. They’re metaphorically more like Einstein’s theory of special relativity, they’re just part of how things work. Cognitive biases are fundamentally tied up in how your brain functions – if you’re conscious, they’re operating.

You can consider you’re brain to operate like a web of information. Any one unit of information sits connected, in context, to other relevant units of information. Accessing one unit of information will more naturally lead to the ability to access it’s surrounding units of information. But this web is built out of your experiences, your understanding, your knowledge, so if you’ve never taken the time to fully flesh out ideas and concepts you don’t agree with then they won’t exist in your web of information and you won’t be able to identify all contextual information even if it’s highly relevant.

Just reading opposing arguments is pointless

Simply exposing yourself to alternative views to the one you agree with is not going to get you very far if you don’t approach the information with a genuinely open mind. Another form of cognitive bias germane to this point is the ‘backfire effect‘ which was displayed when information, which showed that vaccines are not harmful, was presented to people who believed they were. The information made the participants in the experiment double down on their view, and reject vaccines even more staunchly.

In many ways this is the essence of tribalism. People react as though ‘their side’ is under attack, the quality of the information and its implications becomes unimportant and the act of rejecting it becomes a display of solidarity and loyalty with one’s own side. In other words, people get defensive. The backfire effect is doubly important, not only to recognise in oneself, but also in informing how to approach a discussion with someone who does not agree. Simply bombarding them with facts and information isn’t going to work, just as if they did the same to you it probably wouldn’t work. A higher bar must be overcome – even the most detestable opposition must be made to feel as though they’re genuinely being listened to and understood, and an alternative viewpoint should be presented which is sympathetic to their priorities and is built out of relatable points.

Don’t just pick any opposition

Any given argument has multiple versions. The pro-gun argument has a sheer Second Amendment version, a version based on a civilian’s right to defend themselves, a version based on the necessity to maintain the means to overthrow the government by force if necessary, and others. It’s firstly important to take these arguments in turn and be able to defeat them in and of themselves, and together.

But, more importantly, every argument has its worst version. The pro-gun argument has the “the best way to stop a bad guy with a gun is a good guy with a gun” argument. It’s an argument made by the pro-gun lobby, and – without getting into the detail now – it’s fairly illogical. But attempting to set the worst version of the argument up as the one to beat, and then beating it, is both too easy and a mechanism for relegating your own argument to its level, or elevating it to yours. Neither is desirable. Opposing an argument is better done when taking it at its highest point and proving that it’s possible to defeat the best possible version of the argument from whence a defeat of the worst version can flow.

Alas, the ease of picking the worst version of an argument is often too tempting. It is flattering – in a slightly bullying way – to assume alternative arguments simply aren’t of the same calibre as one’s own, and all too convenient to paint them as such. But doing so should not be confused with anything more than politicking, demagoguery or mere incompetence.

This has become ever more pertinent recently as it is particularly easy to get swept up in a Twitter storm and assume those who are most vocal are purveyors of the most prevalent view. Many an Op-Ed has been written to refute a bad argument which, sat in full perspective, pretty much everyone already disagrees with. There might be a handful of people on Twitter or Reddit who the Op-Ed is arguing against, but if an argument appears in a paper of record those reading it may well assume that it’s opposing a widely held view, and the implicit effect of the piece is to convince everyone that the opposition to their righteous stance is fierce and popular. Such an effect is thoroughly counter-productive when the view the Op-Ed opposes is actually rather insignificant. Sending everyone into flushes of righteous indignation is likely to do no more than bring out their least conciliatory side, the side most conducive to entrenchment and division and anger, which may often be useful, but not when the response is unjustifiably larger than the opposition it claims as its raison d’etre.

It would be worth an extra question from editors: who are you arguing against?

Identifying premise

Too often it’s possible to oppose an argument without realising that the alternative arguments come from very different places. The immigration debate is a good example of this. On the left, it’s all too easy to point out that immigrants are net contributors to society who are less likely to be the agents of crime. But to make this argument to anti-immigration proponents is to miss the premise of their argument, which was beautifully described in a two-part report from This American Life.

To people who have seen their wages stagnate as food and utility prices rise in correlation with an increase in the number of immigrants in their community, the points seem connected. The underlying problem is one of financial security. This is widely recognised and yet it’s still too tempting for some to embark on an argument which simply disagrees with the superficial “jobs and crime” point, and fails to empathise with one’s country-folk and thus fails to provide an engaging counter-point.

It’s easy to argue from your own premise, but to a proponent of an opposing argument, your premise is often a complete irrelevance. And while you may have all the justification in the world, that justification shouldn’t be confused with any form of argumentative empathy. Cognitive biases hinder the ability to engage in an alternative understanding, but not doing someone the decency of understanding their priorities and context is to invite the same in return to the detriment of progress.

Facebook’s algorithms are a symptom, not a problem

Facebook has occupied the fulcrum of much of the news and commentary circling news bias, echo chambers, and opinion bubbles. But to blame Facebook is to take too little responsibility for one’s own opinion on the one hand and to give Facebook too much responsibility for same on the other.

One very simple observation must be that Facebook doesn’t create any of the content it circulates. If poorly argued content – the result of the writer’s or editor’s cognitive biases, devised from a misunderstood premise and against a relatively minor opposition – was never created then Facebook would not be able to present it to anyone. Equally, if the reverse were true and the content existed and Facebook did not, the content would still be poor. And one needn’t wallow in the depths of fake news to find arguments created by cognitive biases.

In addition, once Facebook take’s responsibility for one piece of content, according to US law, it is then designated a media company and responsible for all content. Asking for Facebook to deal with our cognitive biases is necessarily asking them to decide everything we see on Facebook, and everything it’s possible to post on Facebook. One can’t overstate the case here, it is a request for Big Brother.

It might be healthier for national discourse for media outlets to educate their writers and readership on cognitive biases rather than stoke anger against Facebook, even if it does make for better headlines.

The Google effect

The way people use Google’s search engine lies at the heart of cognitive biases, and confirmation bias in particular. Google’s search engine is now so good, so effective, and there is so much information now online that anyone writing any article on any topic will almost definitely be able to find supporting ‘evidence’. Just because supporting evidence exists does not mean it’s conclusive, it must be read in context of the other information available, much of which will be contradictory, and while fully cognizant of how it came to be.

The reason this lies at the heart of the issue is because it dovetails two points. On the one hand the ability or propensity of someone to seek out disconfirming evidence. And on the other their ability to analyse the limitations of supporting evidence.

It’s just too easy, for anyone researching anything, to Google a phrase and launch headlong into vindicating evidence without a second thought to an alternative view. It feels good to be vindicated. And so one naturally carries on, unobstructed by the better angels of their nature, into a self-indulgent whirlpool of self-supporting self-understanding. Having furnished oneself with more information on a topic, through inaction, and with no ill will, it’s possible – probable, even – to find oneself in effect less informed on the issue one set out to research.

Aversion tactics

Again, this may not be an exhaustive or orthodox list, but there are certain elements to the process of counter-bias tactics.

Decide that it’s the type of person you want to be

This is the most important element and one which often gets missed from much of the advice. It’s not the kind of thing that’s only available to a few people. But equally no one is saying you have to be the type of person who’s capable of mitigating their cognitive biases, but if you don’t want to be then don’t pretend to be. But note that any efforts can’t just be token, can’t simply be for the sake of a mention at a dinner party table. They have to be heartfelt and genuine otherwise they will be untrue, hollow and ineffective. And if you decide it’s not for you then you may well still be taken by the urge to opine on topics as though you are furnished with the truth, just know that your doing so is disingenuous and self-indulgent.

Fully form an alternative understanding

Christopher Hitchens often spoke of the working philosophies of the Oxford Union. During debates it was, he said, expected that participants could if necessary take on the opposition position and “wear it as though it were your own”. Seen as a mark of intellectual agility, this also acts as a sort of test that could easily be self-imposed. More than merely being aware of alternative arguments, it’s an entirely different thing to see them as a potentially coherent world view. To be able to arrange the context, assumptions and premise around the arguments to form an full understanding.

This is crucial to keeping cognitive biases in check because it generates an ability to access an alternative world, to step in to it and feel it. In pursuit of researching a competing idea one of the most difficult elements is knowing what search terms to use; being host to only one understanding can mean the avenues to pursue another aren’t always forthcoming. But by engaging in the thought experiment of assuming a different position it is possible to chip away at finding material, and often the first chip is really all that’s required to get into it.

Seek to confuse tech algorithms

It’s not just Facebook, two people of demonstrably different political leanings can yield different search results from the same search term. So it’s perfectly possible to notice how many alternative arguments appear in your news feed, newsletters or news apps. The more you search for and click on a range of views the more balanced these results will become. Don’t dismiss that weird friend who posts distasteful views, read what they post and ask yourself whether you can genuinely rebut the point or whether you experience no more than moral indignation.

Test assumptions

This doesn’t have to be extensive, no one is expecting lab results. But, for example, around the special senate race in Alabama late last year there was a lot of talk about voter registration laws which would obstruct the Left’s vote. As a disclaimer, voter registration is something to be opposed on principle alone. But when referring to its real world effect, that’s something which should be measurable and not hard to find information on. A quick search for “impact voter registration” throws up a number of studies including two informative explainer articles from Vox, “Voter suppression in Alabama: what’s true and what’s not” and “A major study finding that voter ID laws hurt minorities isn’t standing up well under scrutiny“.

The first describes some of the false claims that had been made about how restrictive voter ID laws are in Alabama and the second that their effect isn’t necessarily as big as is often claimed. Again, it should be opposed on principle alone. But this quick test reveals that actually voter ID laws aren’t the main problem for Democrats in Alabama. When combined with Democrat voter turn out it becomes clear that the main problem is voter apathy – possibly fuelled by a lack of funding from the DNC – and it is that which won the day for the Dems.

To test assumptions in this case leads to a curious reframing of the situation.

Dennett’s 4 points for critical commentary

Daniel Dennett, the American philosopher and cognitive scientist, famously composed 7 tools for critical thinking which are well worth a read. But also composed four points for countering an argument. These are designed to be used mid-debate but are nonetheless useful to apply to oneself in contemplative moments because more generally they’re used to guard against cognitive biases and some of the less useful effects of the adversarial nature of discussion:

1. Attempt to re-express your target’s position so clearly, vividly and fairly that your target says: “Thanks, I wish I’d thought of putting it that way.”

2. List any points of agreement (especially if they are not matters of general or widespread agreement).

3. Mention anything you have learned from your target.

4. Only then are you permitted to say so much as a word of rebuttal or criticism.

Previous ArticleNext Article
Ben Allen is a traveller, a millennial and a Brit. He worked in the London startup world for a while but really prefers commenting on it than working in it. He has huge faith in the tech industry and enjoys talking and writing about the social issues inherent in its development. Find him on Twitter @benjijamesallen

Leave a Reply

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