Information Cascades in the Election
This past election year was one of the most memorable and insane elections in history, which saw Donald Trump elected to be President of the United States. However once the election was over, half of the country was infuriated and people began to evaluate how one candidate won over the other. One unique aspect of this election was how prevalent social media was. Both Facebook and Twitter were actively used by millions of Americans, and even the candidates themselves, to help their side. The enormous issue brought up after the election was whether fake news articles circulated via Facebook was responsible for swinging the election in Donald Trump’s favor.
Many of these news articles are made up by fake news sites, which make themselves sound like legitimate companies with names like the Denver Guardian. Some of the articles that spread across Facebook were about subjects like how the Pope endorsed Donald Trump and how a Democrat who was going to testify against Hillary Clinton was murdered. Many people have blamed Facebook for allowing these news stories to circulate. However, a concept about networks can be used to explain this situation. That concept is information cascades. If there is an article that hurts Clinton’s campaign, whether it be true or false, it catches the attention of a pro Trump supporter and vice versa. Facebook is a platform based on people sharing information within their networks. Therefore if a Trump supporter sees an article that’s anti-Clinton, they will be incredibly interested in it and they will want to share this information with the rest of their network. To see how this could hypothetically influence an entire election, it’s to best to look at this through an example. If one Trump supporter sees an article saying that the Pope has endorsed their candidate, they could share this article to their entire network of (conservatively) 150 friends. If half of these people also agree with the article, then they will also share the article with the rest of their network. This is how an information cascade is begun, and eventually this news article will reach millions of people even though it isn’t true.
In general this is an incredibly tricky issue to solve, because this type of sharing and engagement is what Facebook’s entire business is built off of. Therefore, they have tried to save themselves from any blame. One argument is that although information cascades can spread quickly, even they are limited after a point. For a person to share an article, they have to be incredibly interested in it. These types of articles are only interesting to those who want to believe them. Therefore if this type of article were to reach a Clinton supporter, it is incredibly unlikely that they would either believe it or spread it further. This means that fake pro Trump articles really only get spread to or affect people who are already pro Trump, and consequently don’t really affect the outcome of a national election. Another issue that Facebook and Mark Zuckerburg have brought up is that they are not responsible for, nor should they filter out information that people want to share because it would impose on an individual’s freedoms and compromise the integrity of the business. It will be incredibly interesting to see if and how the concern over fake news affects social media in the future.
http://www.vox.com/new-money/2016/11/16/13637310/facebook-fake-news-explained
https://www.theguardian.com/technology/2016/nov/15/facebook-fake-news-us-election-trump-clinton