No, I’m not talking about the 2016 or 2020 polls! Lord knows enough people are doing that right now.
I’m talking about 1936, when incumbent Franklin D. Roosevelt took on Governor of Kansas Alf Landon in a bid for the American presidency.
At that time, one of the most trusted American polls was run by the Literary Digest, which was a popular general interest magazine that had been running since 1890. They had correctly predicted all five of the presidential contests preceding that of 1936.
This very trustworthy poll boldly predicted that Landon was going to win 57 percent of the vote and defeat Roosevelt. What actually happened was that Landon got only 37 percent of the vote and suffered one of the worst defeats in American presidential-election history.
In hindsight, the Literary Digest’s error was simple.
They had blasted out 10 million ballots, which was an enormous sample size at the time, but they had made a major error in how they located those 10 million.
They had sent all their ballots out to subscribers to their magazine, people who owned cars (they found them through their registrations), and people who owned telephones. This may have been fine at any other time, but 1936 was in the midst of the Great Depression. The people who subscribed to magazines and owned cars and telephones were mostly people from the wealthier tiers of American society, and at that time the wealthy people were mostly Republicans.
The inevitable result was that they ended up unknowingly asking far more Republicans than Democrats about who they were going to vote for.
So it was that the Literary Digest proved to be disastrously incorrect. Roosevelt won 98.5 percent of the electoral votes in one of the biggest presidential landslides in history.
The error destroyed the Digest’s reputation. It folded a year-and-a-half later.
In a way, the Digest did America a great service. Their error was so egregious and so famous that it changed polling forever. New methods were developed to increase accuracy by sampling more carefully.
Similar transformations may result if it turns out that there’s truth to the current rumor of “shy Trump voters” who are deliberately lying to pollsters. If that’s true, new methods may have to be developed. Or, perhaps the other polling firms will simply have to figure out the Trafalgar Group’s secret.
Then again, maybe Trafalgar proves to be “crazy” as some have said and those stories of shy Trump voters turn out to be all hot air. I’d like to say we have one more day to find out, but it seems more and more likely that it may be weeks or months.
If so, that should certainly be fun (insert sarcasm emoji here).
Copyright 2020 Jeff Suwak