The buzz of the U.S. Elections is all around and the results are much eagerly awaited. The outcome will have a global impact affecting almost every economy due to its jillion of relations with other countries. How beneficial it would be if these results could be predicted and that too with the help of the nature of tweets by the contesting candidates?
In the U.S. midterm elections of 2018, an analysis of tweets from candidates running for Senate leading up to election date revealed that Democrats, who won 2018 midterms were more negative than Republicans on Twitter. This contradicted the famous rally call of Democrats, first popularised by Michelle Obama in 2016: When they go low, we go high.
Aleszu Bajak, a journalism teacher and Floris Wu, a Master’s student in journalism at Northeastern University discovered this relation from an examination of more than 1,24,000 tweets. They filtered the tweets through a machine learning program which derived an average score based on the segregation of tweets according to positive or negative words. To avoid the problem of false positives and false negatives, they performed multiple tests so that the meaning is recognised in the context in which it is said. Out of the 33 democratic candidates analysed, 19 of them tweeted negatively out of which 15 won, which connotes the correlation between tweets and results.
Although the opposite was true for Republicans, that is, the candidate with positive tweets won, an interesting conclusion was derived from this exercise. It bespeaks the impact social media can create even in politics, so how it should be handled becomes significant in political campaigns too because a small tweet can make the minds of voters tweak. A multitude of interrelationships can be found among variables that may seem extraneous; we just need to have that witty insight. So, what do you think? Who will grab the crown: Democrats or Republicans?
Kriti Vishwakarma
Comments