Skip to main content



Information Cascades and the Spread of True vs. False News

Article Link: http://science.sciencemag.org/content/359/6380/1146

Recently “fake news” has become an important problem that society has started to acknowledge as it becomes more and more prevalent due to new social media technologies. Most discussions of false news have revolved around anecdotes, examples and theoretical ideas on how to tackle the problem rather than performing investigations into why this is occurring and how it spreads. This article tackles this by comparing the differences in the spread of true and false news on the internet, namely through Twitter, utilizing methods to analyze information cascades as well as other statistical methods. The authors found that not only did false news and rumors diffused faster and farther than the truth, false news was more novel than true new. They concluded that there may people may be more likely to share novel information. In order to combat false ideas and rumors people should try using novelty as a tool to spread the truth throughout society. The article also found that humans, not bots, spread false news at a much higher rate. In fact bots spread false and true news at the same rate. This contradicts recent political discussions into foreign interference in US elections. In addition, the authors found that the characteristics of the people spreading the information, i.e. the number of followers, had less of an effect on how information spread than the information’s content. This article is a great start into investigating how “fake news” has become such a problem in modern times and how it behaves in an attempt to combat and control it.

This article utilizes the concept of information cascades recently covered in lectures to describe and analyze how information spreads on Twitter. The authors defined a cascade as an unbroken retweet chain with a single common origin. In total, they captured over 126,000 cascades of rumors spread from 2006 to 2017. The cascades grew as more and more people saw something they wanted to share and retweeted it leading to other people seeing this information and who else retweeted or liked it, and deciding if they wanted to continue this chain. The article analyzed these cascades in a variety of ways including depth, size, breath, as well as topic. By doing this the authors were able to determine that falsehood reached depths much faster than the truth and reached many more unique people and each depth due to the larger breadth. The way information cascades were analyzed and described is very similar to the recent guest lecture discussing information cascades on Facebook by Lada Adamic. Information cascades exist in various forms from Facebook and Twitter, traffic, politics, and even guessing the percentage of difference colored marbles in a bag and understanding all them is essential to understanding and improving our world.

Comments

Leave a Reply

Blogging Calendar

October 2018
M T W T F S S
« Sep   Nov »
1234567
891011121314
15161718192021
22232425262728
293031  

Archives