Biden campaign made decisions on ‘probability’ of a polling error: Annie Duke

Max Zahn with Andy Serwer

Members of the Electoral College concluded the 2020 presidential election earlier this week, but crucial questions remain about a widespread polling error that underestimated support for President Donald Trump — just as a similar error did in 2016.

But skepticism about the polls even before Election Day may have contributed to President-elect Joe Biden’s victory, said behavioral scientist and former poker champion Annie Duke.

In a new interview, Duke said the Biden campaign made strategic choices based on the “probability” that the polling error from four years ago would repeat itself, dedicating significant resources to decisive midwestern states despite surveys that showed sizable leads over Trump in the region.

“I think that what they said was, ‘There's a nonzero probability that this polling error is occurring again,’’ says Duke, the author of a new book, “How to Decide: Simple Tools for Making Better Choices.” “And you can see that their resource allocation actually took that into account.”

“Biden was pulling very far ahead in Pennsylvania, Wisconsin, and Michigan,” Duke adds. “And yet, when you look at his ad spending, that's where most of the spending was — including the PACs — because I think that at that point they were taking into account that there could possibly be a polling error again.”

In the final days of the 2016 campaign, statewide polls gave Democratic nominee Hillary Clinton a lead between 4 and 6 percentage points in Wisconsin and Pennsylvania, but she narrowly lost both states and received criticism for devoting insufficient attention to them.

To avoid Clinton’s fate, the Biden campaign visited crucial midwestern states more frequently and spent more money on advertising in them, the Associated Press reported — even though polls showed the Biden campaign with significant leads in the states.

On the eve of Election Day, a polling average calculated by FiveThirtyEight gave Biden an 8.4 percentage-point lead in Wisconsin and a 4.7 percentage-point lead in Pennsylvania. After the votes were tallied, Biden won both states but by slim margins: an advantage of 0.6 percentage points in Wisconsin and 1.2 percentage points in Pennsylvania.

The decision-making fallacy of ‘resulting’

Duke credited the Biden campaign for resisting a decision-making fallacy she calls “resulting,” in which people retroactively judge decisions based on their outcomes. In this case, the Biden campaign didn’t fixate on what the Clinton defeat revealed about shortcomings of that campaign, but rather noted how a polling error that misguided the Clinton campaign might happen again.

“So [you] learn something new in 2016 — something that [Hillary] Clinton couldn't have known about — but as we go into this we can know about it,” she said.

Annie Duke, a behavioral scientist and former poker champion, appears on "Influencers with Andy Serwer."
Annie Duke, a behavioral scientist and former poker champion, appears on "Influencers with Andy Serwer."

Duke spoke to Yahoo Finance Editor-in-Chief Andy Serwer in an episode of “Influencers with Andy Serwer,” a weekly interview series with leaders in business, politics, and entertainment.

She warned that the human tendency to fixate on results can similarly hinder investing decisions when traders over-interpret a big win or a devastating loss. Investors shouldn’t evaluate decisions solely based on their outcomes, Duke warned, noting the role played by chance and luck.

To illustrate the point, she raised the example of the final seconds of the 2015 Super Bowl, in which the Seattle Seahawks controversially chose to throw rather than run, even though they stood just one yard away from the end zone. The pass was intercepted by the New England Patriots, who went on to win the game.

Criticism of the Seahawks erupted online, but Duke says they had made a good choice.

“When you look at these analyses, none of these talk about process,” Duke says. “None of them talk about what's the probability of the ball getting intercepted or incomplete or being a touchdown pass, [and] how does that compare to the probability that [running back] Marshawn Lynch is actually going to score.”

“And of course, those are all the things that you need to know in order to know whether it was a good decision — not just the result of the play,” she adds.

Read more: