An odd thing happened just yesterday that got me thinking about the way the internet works and about the much-talked about issue of confirmation bias.
On Friday, I posted an article about the link between cancer and contraception. In the post, I start off explaining that hormonal contraception has been shown to be a very positive thing both for individual women and for society in general. I then describe how it has been linked to increased rates of breast cancer, decreased rates of ovarian cancer and increased rates of cervical cancer (although that is mostly because hormonal contraception is often linked to higher rates of HPV). I then conclude the piece saying that understanding more about the link between contraception and cancer can hep us identify cancer preventing strategies that can allow women to enjoy the benefits of hormonal contraception without increasing their cancer risk. For example, using barrier methods such as condoms as well as the pill can prevent HPV infections, thus bringing down the risk for cervical cancer. Participating in regular breast cancer screening programs has been shown to decrease the risk of being diagnosed with late-stage breast cancer (obviously). And so on. A useful analogy for my point would be that hormonal contraception is like driving a car. Getting into the driver’s seat of a car carries an inherent risk of being hurt in an accident – yet most people feel that the benefits of driving outweigh the potential risk and therefore choose to drive. However, there are things that people do to decrease their risk of being hurt – such as wearing a seatbelt, driving sensibly, turning on their headlights in the dark etc. Similarly, hormonal contraception is worth the inherent risk to most women – especially if they take the necessary precautions to decrease the risk of cancer. The bottom line is that it was – I thought – an unequivocally pro-contraception piece.
Imagine my surprise, then, when I realized that the piece in question had been linked to by a very right-wing, religious blog – clearly as supporting evidence that contraception gives you cancer and is therefore a risk to be avoided. Whoever linked the post must have either
- Not have read past the headline, or
- Have read it and just ignored all the (extensive) parts where I explicitly make an evidence-supported case that hormonal contraception is on balance a great thing both for individuals and for society.
Leaving out option 1, which is simply a case of laziness, I am much more interested in option 2. Never has idea of “confirmation bias” been more pervasive in society than in this day and age. This is time of such polarized political opinion, and yet we often see the same source being used to make opposing points, which means both parties have read the same thing and come away with completely different take-home messages.
Confirmation bias is a very real and much studied psychological phenomenon whereby people seek out information that reinforces their way of seeing things and will ignore evidence that contradicts their preconceived notions. There are some truly astounding examples of how people will use confirmation bias to confirm their belief in almost anything (including some pretty hilarious conspiracy theories). To make matters worse, in the wake of the 2016 election, people have become acutely aware of the fact that social media tends to show you news stories that you are probably going to agree with (in the same way that they will show you ads for things you are more likely to buy). In the same way, people are more likely to go look for news sites that will have the kinds of stories they will approve of, thus reinforcing their worldview. In a way, this internet driven news bias means that confirmation bias is more of a problem now than it has ever been.
The question that has been on most people’s minds since the consequences of confirmation bias have become evident has therefore been how to best fight it. As it turns out, psychologists have been trying to figure this one out for decades. People will deliberately avoid situations where they can come across evidence that proves them wrong – and the only way they’ll listen is if completely unequivocal evidence is presented to them, something that is very hard to achieve in a non-controlled experimental setting (i.e. when talking about “real world” situations). Which means I highly doubt that any of the people who clicked on my article from the right-wing religious website changed their stance on women’s health simply because I said so.
The question of how to end or at least improve confirmation bias therefore remains very much open. What do you think? There are no bad ideas. An idea that I particularly like (ironically perhaps because of my background) is to include confirmation bias education in science classes for all ages. People who know to look out for confirmation bias and understand it might be better at avoiding it and will certainly recognize it others. The debate. however, remains open.