Tuesday, November 08, 2016
Alf Landon and tonight's nightmare for the polling industry
I haven't heard anyone mention the name yet, but the election of Donald Trump, the Republican candidate for president, which is becoming more and more likely with the announcement of each new set of results, reminds me of another GOP candidate from long ago- Kansas Gov. Alf Landon, who was crushed by FDR in the election of 1936.
Landon won only one state as Roosevelt cruised to re-election, but he left his mark in another way- his loss changed the political polling process.
The best known poll at the time, the Literary Digest poll, predicted a solid Landon victory. The poll had accurately predicted the winner for the previous two decades, but it was a whopping 19 percent off the final result.
How did that happen?
Literary Digest's sampling came from people with magazine subscriptions and various mailing lists, all of which catered to upper class and upper middle class, totally missing the populist wave that kept Franklin Delano Roosevelt in office.
What happened today was a flashback to that election of 80 years ago. The polls were inaccurate and wildly so and once again overlooked a populist wave that gripped the nation.
I have heard speculation that people are giving fake answers because they are fed up with pollsters. While it is undoubtedly true that people have had their fill of pollsters, that is not the problem. The problem is the change in the way people use technology.
Landlines, though many of us still have them, are rapidly becoming dinosaurs. People have their cellphones with them everywhere they go and their numbers are no longer included in thick directories.
Plus, we have all become so inundated with junk calls, many of them from politicians, who made sure that political callers could not be placed on no-call lists, that a large number of people are either hanging up on the pollsters or seeing an unfamiliar number on caller ID, are simply not taking the calls.
There are also far too many politicians who put so much stock in polls that they forget to do the actual job of campaigning and using common sense.
It reminds me of the educational malpractice that takes place when schools spend so much time taking practice standardized tests (their own form of advance polling) to show them how well they will do on the state test at the end of the year. Often, they take multiple practice tests to get a running idea of how the students are doing, overlooking what basic common sense should tell them- they would do far better on the state tests if they didn't spend so much time trying to predict how they were going to do on them.
Fewer polls from fewer sources would be an excellent way to improve things for 2020, or even for the mid-term elections in 2018, but that is not going to happen.
When it comes down to it, polls serve the same purpose as horoscopes and unfortunately, offer the same reliability.