Information Cascades: How Polls Can Affect Our Democratic Process
https://phys.org/news/2019-11-biased-polls-skew-elections-experimental.html
With the fifth Democratic 2020 presidential debate concluding last night, I began thinking about how much of an affect poll numbers do, and will ultimately have, on the outcome of our presidential election. As a result, I realized the topic of my final blog post, and as such, I will be discussing how Information Cascades affect our democratic system. The article I chose to analyze focuses on how biased polls have the ability to skew elections. The article references a paper published by the University of Southampton School of Social Sciences in which they conducted an experiment on 375 participants to determine the veracity of this claim.
The experiment was conducted by asking participants to vote in a series of simulated elections between two parties, K and J. In each of these elections, they would provide the participants with an unbiased condition where they saw all of the polls, and another where they were provided with a biased condition and were only shown polls were a party K was considerably favored. In the biased election, party K won 80% of the time, and in the unbiased election, party K won 60% of the time. In a second experiment, participants were informed that they were being shown biased polls, but party K still benefited with party K still winning 64% of the time in the biased instances versus 57% in the unbiased instances.
This finding was incredibly interesting to me as I felt it was absolutely demonstrative of a concept described in class regarding information cascades. This has to do with the fact that when people are able to infer information from other people’s choices, this public information may be more powerful than their own private information, and as a result, will cause them to join other people in their choices regardless of basing their decisions solely on their own private information. In the case of political polling, this occurs when people initially choose which candidates they feel are better, polls are published, and then people viewing the polls infer information from those statistics, and sometimes this inferred information takes priority over their personal information. The important distinction, which is described in the textbook is that “individuals in a cascade are imitating the behavior of others, but it is not mindless imitation. Rather, it is the result of drawing rational inference from limited information”. In the case of political polling affecting the outcome of elections, a potential voter may view a poll and may draw rational inferences regarding the quality of a candidate; if a candidate performs poorly in polls, a voter may become more sensitive to the negative qualities of that candidate and may begin to question their policy or morality.
In this age of information, I believe that there needs to be more transparency regarding data that is displayed to the public. As is shown by the article included, polls can have a profound and inconspicuous effect on our democratic process.