Poison Pens 8: the art of confirmation bias

Confirmation bias is, speaking simply, the psychological process where you seek out things that agree with your opinion and dismiss those which do not.

It's most commonly acknowledged among market traders, who seek to be ever alert to the prospect that they could be wrong, and try to remain open to contradicting views. The reason they do this is because if they didn't, they'd quickly lose a lot of money.

But for most of us, confirmation bias isn't something we worry about. We like our own opinions, and we like those who share them, mostly. Sure, we might like people who disagree with us too. But being around them all the time would eventually become... disagreeable. So we don't.

Confirmation bias is one of the things that makes people interact with the media they choose. It's a long-standing canard that tabloid readers buy the redtops in order to confirm their own neanderthal point of view on the world.

In fact, it's much more likely that the very people levelling this argument, smug broadsheet readers, are themselves the biggest victims of confirmation bias.

Tabloid readers are very aware that the world consists of more than X Factor, celeb affairs and premiership football. They read them like some women read Hello! or Elle - for the escapism away from the humdrum reality.

Broadsheet readers, however, do not have the same level of disconnect with their chosen media. Let's take, as an example, the following piece from today's Observer, entitled 'US shaken by sudden surge of violence against gays.'

According to the title, there is a surge of anti-gay violence sweeping across America, which is a shocking prospect, not least to the gay-friendly readers of the Observer.

Yet, the evidence doesn't remotely support that. Within the article, the entire evidence offered consists of one suicide (well-reported, and subject to legal action towards alleged bullies), one quite vicious beating of a gay man and some vaguely euphemised 'other youths' in New York, a group of men who had a bin thrown at them, and a customer of the Stonewall Inn who was robbed.

I don't wish to diminish the seriousness of any of those individual acts, but frankly, all of that and much worse likely happened in any single Dublin housing estate last night. It doesn't add up to anything close to a surge of violence against gays in a country of 300 million people.

The article opens with the poignant scene of one brave man's 'lonely vigil' protesting for civil rights for gay people. Then swiftly moves on to inform us that, like him, 'Liberal America looks on aghast as virulent homophobic prejudice seems to have returned to its streets and cities.'

'Seems' is the key word here, the mealymouthed weasel word which excuses the writer from offering objective fact and allows him to present biased opinion in its place. It's a technique pioneered and perfected by the dark minions of the Daily Mail, where insinuating invective has been house style for many years.

'Seems', 'may be considered to be', 'is now thought by many' - such phrases are bullshit euphemisms. They are shorthand, telling the reader - here is what you should conclude, here is what you should be thinking.

Ironically, the very people who would be impervious to this technique in the Daily Mail are suckers for it in their own favoured papers, like the Observer.

This is confirmation bias in action.

The issue isn't whether there is a tidal wave of anti-gay violence in America or not. Clearly there isn't, and even the article is forced to concede in a single hurried line that the US is set to overturn the ban on gays in the military and Florida has just introduced gay adoption.

The issue is that the Observer, for some reason, has thought it legitimate to pretend that there is, using all the Daily Mail's rhetorical tropes of insinuation to depict skewed opinion in place of objective fact, when those facts do not add up to support their argument.

Journalism is becoming ever more debased, and nowhere is this more evident than in the broadsheets. Don't let confirmation bias blind you to that.