Wednesday, May 4, 2016

Is it possible to predict the next US president accurately


This year is a special year because a new president will be chosen to lead the US. With that being said, this is always an interesting time to follow how people use different methods to make predictions as to whom will be elected in the fall. In this article published back in November of 2015, the authors mention Western Illinois University who predict Bernie Sanders winning the election. They argue that the system which was used has been able to predict every presidential election since 1975 with 100% accuracy. In their report, Bernie Sanders is predicted to win and by a massive landslide. It was said that Hillary Clinton survives past Super Tuesday, but loses out before the month of March is concluded.



With that being said, now that we are in full swing we are able to take a step back and see how some of these tools were not a good predictor. We are currently at the point were Hillary is still in the race and is actually winning. She is leading with 1682 delegates while Bernie has 1361. This is definitely not what the system predicted. Based on this, would it be possible to make a better prediction now from the pattern we have thus far? What are the key parameters that are taken into account to make the prediction? 

Overall, it would be really interesting to get some feedback from group and see if they can develop a better system based on this major upset they did not predict.


3 comments:

  1. I agree that this questions is particularly important, specially because grant funding and so many scientific jobs will depend on who wins the presidential elections. However, I believe that it is extremely difficult to predict some factors. For example, Donald Trump started the race with no political experience and using his "own money". In fact back in June he was polled to have the lowest percentage support for the nominations. Could we mathematically predict the public response to his comments like the ones in the following link (http://cdn.static-economist.com/sites/default/files/images/2015/12/blogs/graphic-detail/20151212_woc733_2.png) I believe not. However, people still attempt to use statistics and data to determine the outcome of the presidential elections. Nate Silver in his FiveThirtyEight website attempts to do exactly that, but people could claim he is failing since some of his predictors can't take into account the unpredictable comments from Donald Trump or the many of the other candidates.

    ReplyDelete
  2. I prefer Sam Wang's method. It's all about the mthods, btw.
    http://election.princeton.edu

    ReplyDelete
  3. I prefer Sam Wang's method. It's all about the mthods, btw.
    http://election.princeton.edu

    ReplyDelete