Election Polls Oversight: A reflection on Forecast Accuracy and Bias[shortposting]
The media, locally and internationally, has been labeling this Election
results as Unexpected, based on early predictions that did not realize.
We kept repeating in this class at various times that we should not use
software or mathematical model as-is without injecting our knowledge
of our environment, and resorting to tweaking few things to account for
that knowledge and for possibly intangible, non-quantifiable factors.
1) Discuss what you believe were the flaws in the Predictions made for
the Election 2016? Was it due to? The Method: Was it the way the polls were collected? Did they reach out to the
full spectrum of the population? (State, gender, age, education, occupation,
etc.)? How about the so-called sophisticated forecasting and statistical models
that were developed by Experts?
Bias? Was there a bias in the entities in charge of the Polls
Human factor: Could statistical sophisticated models capture human feelings
such as resentment, aspirations, fear, etc.
Other factors? Please indicate 2) What are the lessons learned from this? If you were a consultant in
charge of the next presidential Elections predictions, what would you
make sure to account for in your forecast?
Rules: This is NOT meant to turn into a political forum. Please refrain from
politically charged statements or any blatant endorsement of either
candidate. Granted the analysis most likely be touching on economic,
cultural, and political factors, but the intent of this exercise is to focus
on the PROCESS of prediction, and not expressing endorsement or lack
of towards any of the candidates.
Failure to abide by these rules will result in dismissing the assignment. Requirements:
Answer question 1 and 2 above by expanding on them
Include at least 2 sources (internet) in your discussion and arguments
Report should be a minimum of 3 pages and max of 5 pages