There is an extremely low probability of www.fivethirtyeight.com, www.electoral-vote.com, www.dailykos.com/blogs/Elections, (the greatest Statisticians, and Data Scientists), and www.predictit.org, (persons with money on the elections), having all been wrong.
The probability that the above Election Prediction websites were all wrong, (about 6% or less), is so low that it is impossible that the results of the election could have actually occurred.
Most persons are familiar with Election Prediction websites combining possibly biased polls, in order to eliminate bias. However, Data Scientists also have the capability to correlate "independent" (or "explanatory") variables, in order to determine the behavior of a "dependent" (or "response") variable. The reliability of these correlations of independent variables to election performance are verifiable for maximum confidence. All the correlations of winning an election having failed in the 2016 presidential elections, is probably not possible.
On election day, Donald Trump voters were advised to ignore polling services, election predictions, and exit polls. The reason is obvious, election fraud had been planned, and was being committed on a massive scale by the Republicans, or an organization acting on their behalf.
The possibility of election fraud can not be ruled out. Increases in Republican voters, and decreases of Democrat voters, are also explainable by the possibility of election fraud. It is not very plausible that Confirmation Bias explains the dramatic failure of a multitude of Data Science-based election predictions. Another verification of the viability of election fraud having occurred is the phenomena of adjusting polling methods, and statistical inference methods, to produce predictions of Trump winning, and thereby possibly creating less accurate future predictions, (because of inaccurate prediction models).
The probability that the above Election Prediction websites were all wrong, (about 6% or less), is so low that it is impossible that the results of the election could have actually occurred.
Most persons are familiar with Election Prediction websites combining possibly biased polls, in order to eliminate bias. However, Data Scientists also have the capability to correlate "independent" (or "explanatory") variables, in order to determine the behavior of a "dependent" (or "response") variable. The reliability of these correlations of independent variables to election performance are verifiable for maximum confidence. All the correlations of winning an election having failed in the 2016 presidential elections, is probably not possible.
On election day, Donald Trump voters were advised to ignore polling services, election predictions, and exit polls. The reason is obvious, election fraud had been planned, and was being committed on a massive scale by the Republicans, or an organization acting on their behalf.
The possibility of election fraud can not be ruled out. Increases in Republican voters, and decreases of Democrat voters, are also explainable by the possibility of election fraud. It is not very plausible that Confirmation Bias explains the dramatic failure of a multitude of Data Science-based election predictions. Another verification of the viability of election fraud having occurred is the phenomena of adjusting polling methods, and statistical inference methods, to produce predictions of Trump winning, and thereby possibly creating less accurate future predictions, (because of inaccurate prediction models).
Specific examples of the improbability of all election prediction websites failing:
See the election prediction graphs below, (the first graph is the most eerily unexplainable):
No comments:
Post a Comment