How We Can Better Predict Election Results Using Simple Statistics
Using Brexit Poll Data To Predict The Referendum Result
Each election cycle, the public is treated to a slew of election polling. Conducted by a myriad of organizations, these polls attempt to represent public opinion regarding specific issues. Increasingly however, there has been a politicization of polls which have been placed in the same basket of “fake news”.
In fact, News agencies often cover how polls failed to predict the election result. With the Brexit referendum for example, the Washington Post, Guardian, Telegraph and CNBC have all covered how polls and pundits where wrong about the election results.
This Article attempts to show how polling is simply a measure of a small part of the populations at one specific point in time at a specific position. It will also show why these ‘specific’ conditions are important. The analysis will be done on RStudio. For Reference the entire code used is up on my personal GitHub.
The Image in the above shows how the data frame of Brexit data looks. The Brexit data is built in to the dslabs package. It contains all of the polls released in the UK from six months before the election. It shows the proportions of voters who said that they would vote to ‘remain’, ‘leave’ and were…