Land of uncertainty.

Photographer: Darren Hauck/Getty Images

Don't Complain That the Polls Were Wrong

Narayana Kocherlakota is a Bloomberg View columnist. He is a professor of economics at the University of Rochester and was president of the Federal Reserve Bank of Minneapolis from 2009 to 2015.
Read More.
a | A

One big complaint about the latest U.S. election is that “the polls were wrong.” It reveals a complete misunderstanding of what polls -- and statistical forecasts in general -- are supposed to do.

Many people confuse forecasts with fortune-telling. A fortune teller seeks to predict the future with certainty -- an endeavor that I view with suspicion. Statisticians use available information to help us make better decisions when we can’t be sure what will happen in the future. This is essential.

Let’s use a sports analogy. Suppose a football coach has to decide whether to attempt a one-point or two-point conversion after a touchdown. He can’t know with certainty which approach will be successful. He can, however, know how successful they have been in different situations in the past. He can even ask a statistician to calculate probabilities based on a simulation of millions and millions of attempts. He would then combine that information with an assessment of his current situation to make an informed choice.

Polls play the same role. Suppose a candidate must decide how to divide ad spending between New York and Iowa. Again, there’s no way to know for certain which approach will generate the most votes. A poll gauges what the candidate’s average performance would be if the election were held millions and millions of times -- an average that we know will probably not describe this particular election. Nonetheless, as a guide to decision-making, it’s useful to know how many votes the candidate is likely to get in each state (if the election were repeated many times).

Most of us probably aren’t making decisions that hinge on the outcome of the election, so for us the polls are just one more piece of news. They matter a lot more for candidates, and some poll or polls presumably informed Hillary Clinton’s decision, for example, not to campaign personally in Wisconsin. That might have been a mistake, given what we now know. But we’ll never know whether it was wrong given the information that the pollster and Clinton had at the time. 

How the Polls Got It Wrong

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

To contact the author of this story:
Narayana Kocherlakota at nkocherlako1@bloomberg.net

To contact the editor responsible for this story:
Mark Whitehouse at mwhitehouse1@bloomberg.net