Skip to main content



The Majority Illusion

The majority illusion is a phenomenon that occurs when the minority in a group is better connected than the other members of the group, or simply more vocal about their opinions. This can be seen in the first graph in the article, which illustrates an example town. In the graph, only 3 nodes were in support of baseball caps where as 11 were against them. The structure of the edges in the graph, however, helped to illustrate the majority illusion. Since the 3 people who were in favor of the caps had far more edges connected to them, the majority of nodes found that the majority of their friends were in support of baseball caps! This rather surprising result shows how the influence of someone’s opinion in a network is dependent on how well connected the node is.

This idea can be used to explain a variety of phenomenon that occur in various networks. One example stated in the article was how public opinion can be controlled by a small, vocal group. People tend to base their opinions on the opinions of others, so if you consider the example with the baseball caps, the people who were unsure of whether or not they liked baseball caps may have said they do simply because they thought it was the majority’s opinion. The article details the same-sex marriage polling, and claims that the majority issue was at work.

I thought that this article was an interesting way to take the graph theory we learned in class and use a phenomenon present in certain graphs to try and describe real life trends. The article also outlines the “friendship paradox”, which can be proved in rigor, but is essentially a generalization of the findings from the example baseball cap town: your friends always have more friends than you, on average. This idea has many applications as well.

Source: http://www.washingtonpost.com/graphics/business/wonkblog/majority-illusion/

Comments

Leave a Reply

Blogging Calendar

October 2015
M T W T F S S
 1234
567891011
12131415161718
19202122232425
262728293031  

Archives