Summary: By nightfall after Trump’s election we had rationalized away this astonishing event with myths, ensuring that we learn nothing from Campaign 2016, Let’s strip away these pleasing stories and confront the truth — no matter how painful. Then we can start to prepare for a better choice in 2020.
- Myth: Trump won the election!
- Myth: The polls were very wrong!
- Myth: The election repudiated US elites!
- Myth: 2016 was a victory for populism!
- The big lesson from Campaign 2016.
- For More Information.
(1) Myth: Trump won the election!
The final counts are not in, but Clinton currently leads by over 200,000 votes (roughly 0.2%). Estimates for her actual total are well over a million votes. The Electoral College put Trump in the White House, as it did with Bush Jr. in 2000 (Gore won by 440 thousand votes, 0.5%.
Like our bizarrely allocation of votes in the Senate (our version of Britain’s “rotten boroughs“), the Electoral College is a historical artifact of our governing system that has outlived its utility — but we are too lazy to fix. For more about it see this article by Scott Lemieux (Prof of Political Science, College of St. Rose) in The New Republic.
(2) Myth: the polls were very wrong!
People saying the polls did not predict correctly predict the outcome often point to the surveys in the months before the vote, forgetting what polls do. First, polls measure the public’s current intentions — they don’t predict future votes. The correct comparison is between pre-election surveys and the actual vote. The last average of 4-way polls tracked by RealClearPolitics gave Clinton a 3.3% lead — vs. her current lead of 0.2%. The remaining uncounted votes will narrow this gap, perhaps giving her much larger win. (California has 4.4 million of them; if she gets 62%, that’s 2.7 million more votes.)
Second, polls are statistical tools — not Dr. Strange’s magic spells. They should be presented with error bars showing the uncertainty of the survey. News stories about election polls and forecasts don’t have them, nor do stories about economic surveys, nor do stories about climate data and forecasts. The press want simple stories. Experts who insist about discussing uncertainty get no more calls from journalists. We’re ignorant because we read the news.
Consider the uncertainties in estimating the vote from polls. Only a small fraction of people answer polls, only some of those will answer honestly, and only some of those who answer will vote. Then there are systemic challenges in any polling methodology, such as extrapolating from the sample who they call and who answers to the actual population of people who vote.
The average of pre-election polls missed the actual result by 3.1% (estimated 3.3% vs. actual +0.2%). I have not found the margin of error for the major polls, let alone the average of polls, but I’ll bet that is within the 95% margins.
Third, our whining about the polls is so modern American. We get the poll accuracy we pay for. Journalists could pool their money and have a few large accurate surveys. But polls are fillers between the advertisements, so each corporate media wants its own. So we get dozens of polls, none very accurate. For more about the margin of error in polls see this note by Pew Research.