Election 2016 Surprise:
The polls will be proven wrong
By Mr. E
[ NOTE: The writer is an Independent, "NPA / No Party Affiliation,"
not a Democrat or Republican, and is not supporting either candidate ]
Almost everyone who has a familiarity with mid-20th century American history knows the story of the 1948 US presidential election, most famously publicized by the inaccurate “Dewey defeats Truman” newspapers, which incidentally carry a high financial value to collectors.
There is no need to go back nearly as far as that however, of course, to prove the difficulties involved in political polling and the resulting inaccuracies, as was demonstrated a bit over a half century later in the US presidential election of 2000.
In every election since then, even when the winner was correctly predicted based on polls, those political pre-election polls that ask for whom the respondent plans to vote have still been quite unreliable, with the actual votes for the candidates often being far higher or lower than rolling political poll numbers from even the day immediately preceding elections had suggested / found.
There are many important technical details that need to be properly addressed in the preparation and carrying out of the hoped-for nationally representative sample survey, and due to many factors the actual final sample can vary widely from being truly representative.
Some of the biggest reasons for this are:
A. Undecided Likely Voters. The percentage of undecided voters in the 2016 election campaign is very large, some estimates place it at over 20 percent.
Even if it is between 15 and 20 percent that is still a very large number that upon deciding who to vote for only on election day itself could sharply cause a large change in the support of the candidates, (at least) enough to prove the polls wrong. This seems very likely for the 2016 election.
B. Economically Poor Voters. In considering poor people who do not have a landline telephone, or a cell phone, and who also do not have an internet connection, the only way for pollsters to ask those people for whom they plan to vote is to send someone to physically knock on their door.
Since no pre-election polling organizations and no polling companies conduct those face-to-face interviews, a growing segment of the population is being excluded from the polls.
Those who dismiss this category as tiny and insignificant will be proven wrong, as the economic squeeze on the 99 percent is causing the number of people in this category to grow.
C. Cellular Wireless / Internet Only Voters. A similar point is that there is a growing percentage of people now who have and use only a cell phone, and do not have a landline at all. Then another group of people has neither a cell phone nor a landline, and has only internet service.
A major result of these developing trends is that most pollsters have gotten more sophisticated over the years to attempt to do whatever possible to minimize (if not eliminate) biases to the survey results, with varying degrees of success.
Random digit dialing, if configured properly to include cell phone numbers and VOIP numbers, makes a large improvement over the years-ago method of using the public phone directory to derive the sample.
However, it still leaves out the small percentage of people who have no phone number of any type, who are not able to be contacted by phone.
Another improvement is that recently some polls are including combined results from internet respondents and telephone respondents, yet that still again leaves out the small percentage with no internet and no phones.
D. The Weather. One only need to go back a few months in time for only the most recent example of this, the "Brexit" vote that took place in England/UK in late June 2016.
All of the polls and pundits proclaimed before the vote that it was close to certain that the result would be a win for the "remain in the EU" choice.
But heavy rains came on the day of the vote, and that combined with undersampling voters in London and major cities who were strongly opposed to remaining were the largest reasons for the most recent failure of the pollsters.
This same possibility should not be discounted whatsoever in the US presidential election of 2016, firstly because of the late date, November 8 instead of the first or second day of November, and if you think that 6 or 7 days doesn't make a difference, you will likely be one of those proven wrong.
In the span of those 6 or 7 days, the weather could significantly worsen, with extreme cold and even snow, especially in heavily-contested states in the geographic "top row," meaning Minnesota, Wisconsin, Michigan, and Illinois, the latter two of which have very frequently (almost always!) been among the very last to decide for whom to vote, and that only on election day itself.
Clinton's small lead in polls in at least 2 of those states is within the margin of error and could easily evaporate and be reversed with extreme weather conditions that are not whatsoever unlikely and that usually cause lower voter turnout in the affected places.
E. Overconfidence. It also could be that supporters of the candidate leading in the polls may decide there is no need to actually go to the polls to vote; they conclude that since their candidate is winning in the polls, they will surely "win" on election day, even without their own personal vote.
The larger the number of haughty and arrogantly-behaving ("deplorable" traits) people that do not actually cast their votes becomes, the more likely it is that the pre-election opinion polls from even the day before the election will be wrong.
This possibility of overconfidence is not unlikely whatsoever, and combined with the previous point "[unpleasant, even severe] weather", could significantly affect the final total of actual votes cast.
One of the ways that many, in fact most, of the polls are inaccurate and misleading is by only asking about (or reporting) support for Trump or Clinton, when significant percentages of voters plan to vote for Johnson, Stein, and even write-in Bernie Sanders.
The mainstream media and many, many people who are too quick to place their complete faith and trust in pre-election political polls are going to receive a huge surprise in the upcoming 2016 election that is drawing ever closer each day.
The polls will be proven wrong again.
You heard it here first.