Post-Truth, Facebook, and the Choice of Information
This month I attended a panel discussion on post-truth with panelists from a wide range of disciplines. Despite our very different perspectives we quickly focused on the role of post-truth in political debates, since even here emotions and impressions have become more important than facts, sometimes with enormous detrimental effects. We all agreed that post-truth was a problem and that more political debate was necessary. A key factor that was identified that evening is the increasing ability to simply ignore "the other side", that is people and opinions we don't like. The debate was inspiring and the importance of the topic makes me believe that I should share a few more insights from mathematical social sciences. This is not because these insights should be thought of as objective truths, but they are rather "food for thought" to foster future debates both inside and outside the academic ivory tower.
Common Sense and Segregation
The economists' theory on Social Learning, that is the process of finding out for instance whether Brexit is good or bad for the British society, is based on the idea that people influence each others' opinions. One of the necessary conditions for finding a consensus is that the social network is connected. This basically means that anyone's opinion can reach everyone else via some chain of other people. This very precise mathematical formulation makes one thing pretty clear: a society that is "too separated" will not reach a consensus. This seems to be a trivial observation and common sense; nevertheless large parts of the current debate focus on dividing societies rather than uniting it — the story of a woman who divorced her husband because he voted for Trump being only one of the many low points on both sides. Yet this story illustrates the problem: if even a couple that has been married for such a long time is not capable of working it out, how should those who live in a "filter bubble", who are surrounded mostly by people of the same opinion?
Facebook and other Forums
Many claim that social media are responsible for the spread of fake news.
But even before Facebook and Twitter many private discussions were based on opinions rather than on facts.
Just a couple of examples from my youth: is the detrimental effect of Marijuana worse than that of tobacco?
Is chewing gum stimulating brain activity and should therefore be allowed during class?
We had many discussions, and the information we used was typically a study that the friend of a friend read about in the newspaper...
But while we were stuck with out teachers and parents whom we were not able to convince, it is now very easy to change the social environment: instead of changing school, sport clubs, or employer we simply adjust our friends list on Faceook. Hence, the mixture of different opinions that ensured discussion and prohibited too extreme positions to be taken in public does no longer exist. Our choice of people and news agencies to follow can be changed with a simple click, at zero cost. If news are unwelcome, we cannot only ignore them, we can ensure that the messenger will not send us anything anymore. Within our bubble, extremist views will suddenly not be condemned but hailed, the more extreme the better. Sharing a good story brings us into the spot light, at least for a few moments. And to be very clear here: this is true for both sides of the discussion. Using a picture from the old Roman city my fellow panelist Teun Dekker pointed out that there are too many "market places", where people decide with whom to interact, and not enough "forums", where all opinions are present. Of course you should spend your time with the people you like, and those typically have similar views on the world. But this is not true for politicians and other persons of public interest. A potentially dangerous move was the decision of publicly funded TV channels in Germany (after pressure from the largest parties) to exclude the right-wing AfD from a TV debate prior to the elections in 2016. These panel discussions are one of the few forums left, and boycotting it is not only a poor but an absolutely counter-productive statement. Ironically this form of exclusion from political debate is exactly what confirms any biases against the establishment.
The segregation of a society into groups with very different and potentially extreme beliefs is one problem that has been amplified (but not been caused) by social media.
But there are other, deeper and more subtle, insights about people's judgement that economics can provide.
Evidence for our inability to interpret probabilities has already been provided in the 70's.
But a recent study (Möbius, Niederle, Niehaus, Rosenblat, 2014) suggests that when people have to guess the probability that they performed well in a test, they overestimate positive signals (i.e. signals that suggest a good performance) and underestimate negative signals.
Interestingly this misjudgement is significantly weaker when they have to guess the probability that someone else performed well.
Earlier studies in a similar flavour showed that people link their own good performance to their abilities, while they try to explain bad performance with their environment.
How can we link these findings to post-truth? Suppose that instead of asking people how good someone performed in a test, we are telling them their result and ask them how difficult they think the test was. Above results would suggest that underperformers will typically assume that the test was hard. Which leads us back to our society: If somebody is unhappy with her or his life and you give two possible explanation, namely one in which he simply performed poor, and one where she or he was disadvantaged, what will the outcome be? So, maybe the success of protest parties does not lie in their "alternative" approach to politics. They rather picture a society in bad shape that has caused our misfortune, thereby offering us an explanation for our struggles that we happily believe. (For instance the USA being exploited by other countries and illegal immigrants, or the European Commission bossing around the British nation.) Next, they offer simple solutions (protectionism and a big wall, or Brexit), which might actually be good solutions if these problems were actually true. (By the way this point has been made by the movie "Our Brand is Crisis".)
There are no easy solutions that could prevent fake news from spreading and our society from getting more separated. Nevertheless the insights above suggest a few measures that can be taken.
It will be impossible to change the way people process information. But we have to remember the shortcomings of the human mind when we communicate with each other. So, instead of asking "How could you believe that...?", we should ask "Why do you believe that...?" We all have reasons to trust one or the other news agency. But typically it has nothing to do with our ability to verify any news, but with our sympathy for one or the other perspective.
There are basically two ways to overcome the problems of post-truth in the flow of information through our society. One involves well informed people and open discussion. The other involves censorship and interferes with one of our highest values, namely the freedom of speech. Assuming that we would like to save the latter, it becomes clear that we all have to become active. There are a few conclusions that can be drawn from the theoretical and empirical findings on information flows, but they all boil down to the ability to discuss.
So, I have decided to change my behavior with respect to news consumption. I do not filter my newsfeed: I allow for news I do not like. If I am able to identify them as fraud, I do so and try to open a discussion. Well connected networks are essential. I share news and contributions only if they come from reliable sources with only little or no self-interest. Once fake news are out there, they will have an effect even long after they have been proven wrong. I do not edit my friends list based on their posts, as this list is the forum where I can foster debates.