Opinion polling is an applied science. Opinion polling as it relates to campaigns for President of the United States is the most visible face of that science, one that sometimes produces a very visible false result. Though watching the polls can be amusing, it is well to remember this tautology: we won’t know the result of this coming Tuesday’s vote until we, as a nation, know the result.
On the day before the election, in November 2016, most polling organizations were confident that a win by the Democratic Party candidate, former Secretary of State Hillary Clinton, was likely. This wasn’t because (as some have suggested) they had their eyes only on the popular vote numbers and the electoral college arithmetic confused them. No: the pollsters were quite familiar with how that arithmetic worked.
The erroneous calls arose largely from the effects of late-breaking developments in that campaign, such as a late announcement by James Comey, then the director of the FBI, that he was re-opening an investigation into Hillary Clinton’s handling of State Department emails during her time heading that department.
The Thing to Know:
Polling is not guesswork or “propaganda.” It is a scientific enterprise. But, like science in general, it produces falsifiable hypotheses. And sometimes that which is falsifiable turns out to be … false.