A View from Emerging Technology from the arXiv
Sandy Hook, the Gun Control Debate, and the Insidious Influence of the Filter Bubble
Do tragedies force us to expand our views on controversial topics such as gun control? Unfortunately not, say Web researchers who have studied surfing habits during America’s worst school shooting.
On December 14, 2012, Adam Lanza shot and killed his mother, then drove to Sandy Hook Elementary School in Newtown, Connecticut and gunned down 20 children and six adult staff members before killing himself.
The incident was the deadliest shooting at a school in U.S. history and triggered an intense debate about gun control. That debate continues today.
One problem is that there is growing evidence online that people tend to seek out views that agree with their own and rarely encounter alternative points of view.
“This so-called ‘filter bubble’ phenomenon has been called out as especially detrimental when it comes to dialogue among people on controversial, emotionally charged topics, such as the labeling of genetically modified food, the right to bear arms, the death penalty, and online privacy,” say Danai Koutra at Carnegie Mellon University in Pittsburgh and a couple of Microsoft researchers, Paul Bennett and Eric Horvitz.
But does that remain true even when serious events occur such as the Sandy Hook incident? Koutra and co decided to find out by analyzing the Web browsing behavior of people who looked at a wide range of gun-related sites and seeing how it changed before and after the massacre.
They began with the data set consisting of anonymous Web browsing logs for more than 25 million users in the U.S. between November and December 2012.
Koutra and co then filtered these logs for nonrelevant queries, such as those to do with video game gun controllers and so on. This left them with 6,900 queries relating to the gun control debate.
They then created a list of over 1,500 websites with related content and classified each as falling into the category of favoring gun control, favoring gun rights or balanced. They also divided the websites in these categories into subsections indicating whether they were purely factual, moderate or extreme in their views.
“From the extracted on-topic web pages 25% are purely factual, 10% highly balanced, 31% and 21% moderate and extreme gun rights respectively and 9% and 4% moderate and extreme gun control,” say Koutra and co.
They then studied the browsing habits of individuals to find out the variety of information that they accessed on this topic. The results give clear evidence of the “filter bubble” at work: most users access a small number of sites, most of which fall into the same category. “All in all, people use the Web to largely access agreeable information,” conclude Koutra and co.
Next, they studied how the pattern of browsing changed after the Sandy Hook shootings. The first thing to note is that after the tragedy, there was a sudden increase in the number of people accessing gun-related websites. But their conclusion is that whatever content people already accessed, the tendency was to continue to access agreeable content but of a more extreme variety.
That’s a rather depressing insight into the Web surfing habits of the American public and an uncomfortable confirmation that our filter bubbles are operating as expected.
There is one potential silver lining. Koutra and co say that with their better understanding of how news events alter surfing habits, it might be possible to devise strategies that expose people to a greater variety of views and therefore to broaden the debate.
Exactly how that could be done isn’t clear. But it must surely be worth some significant research to work how.
Ref: arxiv.org/abs/1405.1486 : Events and Controversies: Influences of a Shocking News Event on Information Seeking
AI is here. Will you lead or follow?
Join us at EmTech Digital 2019.