Statisticians might soon be able to predict the winner of the 2016 presidential election by studying individuals’ activity on Twitter. In fact, it could become the new way of polling. Researchers — including sociologists, statisticians, and political consultants — have found preliminary evidence that social trends on Twitter can accurately predict the results of elections.
Other Twitter-related studies have shed light onto this merging science. In 2010, a study was able to accurately determine the success of a film by studying various aspects of Twitter. The next year, a German student associated the performance of the stock market with particular Twitter variables. It is only logical to believe that human elements within society — elements that determine the outcome of elections, the success of a movie, or the stock market — can be analyzed and understood through social interactions online.
An Indiana University study found that there was a direct correlation between the numbers of Twitter mentions of a member of the House of Representatives candidate and his or her results in the election. The researchers took a random sample of 542,969 tweets that referred to candidates before the 2010 elections. Out of the 406 races, 404 of the winning candidates were mentioned the most on Twitter, compared to their opponents.
Surprisingly, it didn’t matter if the references to the candidate were positive or negative. The fact that more people were talking about the candidate was a good sign. “If people must talk about you, even in negative ways, it is a signal that a candidate is on the verge of victory,” said sociologist Fabio Rojas.
To guard against skewed results, the study took into account two factors. First, that if one of the candidates was an incumbent, he or she was not considered because he or she would already be talked about more on Twitter. Second, candidates that were unequally discussed on media outlets would be eliminated from the sample. “From the beginning, we were looking to construct simple and easy to operationalize measures of political behavior and attitudes that could be useful to social scientists,” said Joe DiGrazia, one of the researchers. Only 8% of adult Americans are said to use Twitter in their daily lives but the researchers maintains that their computations and study have tangible results.
As Internet access is widely available, people are able to express their opinions to a wider audience. The next decade could see the creation of many social media-based analytic polls that predict outcomes by directly evaluating user information and opinion. The first, large-scale implementation of this method could be used during the 2016 presidential elections. Polling presents a plethora of issues, such as sampling bias and bias created from the certain wording of a question. While there are professionals who have done their best to perfect this science, using Twitter to measure voter opinion could be a much more precise measurement.