DONALD TRUMP’S presidential job-approval ratings have been the source of a puzzling paradox. At various points in the election cycle, the share of voters who approve of the president’s performance has been higher than his share of the vote in pre-election polls. According to FiveThirtyEight, a data-journalism website, 43% of voters approve of the job Mr Trump is doing as president, but only 42% say they’ll vote for him. This slight mismatch has prompted election analysts to wonder whether this might be evidence of the ever-elusive, so-called “shy Trump” voters.
The logic of this argument is appealing. If more voters approve of Mr Trump than are willing to say they’ll vote for him, then they could simply be lying to pollsters about their intentions at the ballot box. However, this omits one crucial point about the ways that polls are conducted. It also ignores polling data on which way undecided voters actually lean.
The primary issue is that respondents have more choices in answering questions about who they’ll vote for (what pollsters call the “ballot test”) than they do when asked whether they approve of the president. In YouGov’s most recent poll for The Economist, voters were given five options for the ballot test: they could vote for Mr Trump, Joe Biden or a third party, or signal that they were unsure about their choice or were not likely to vote at all. But when asked about the president’s performance, they had only three: approve, disapprove or unsure.
Faced with more choice in the ballot test, voters are less likely to select one of the two main options than they are in the approval-rating question. This is evidenced not only by the gap between Mr Trump’s approval rating and his vote share, but also by a persistent difference between his disapproval rating and Mr Biden’s vote share. According to FiveThirtyEight, 54% of voters currently disapprove of the president, while 52% say they will vote for Mr Biden.
This calls into question the idea that Mr Trump has a hidden trove of support among apparently undecided voters. If the president’s approval rating is supposed to be a guide to how they will cast their ballots, then surely the same goes for his disapproval rating. And by that logic, there seem to be at least as many hidden Biden voters as Trump voters.
However, we don’t have to rely on these aggregate trends to guess how undecided voters lean. Using YouGov’s polling data, we can also use the president’s approval rating among this specific group of voters to predict how many might eventually support him in the election.
According to the most recent poll, conducted between October 11th and 13th, only 31% of undecided voters approve of the president—12 points lower than his rating among all likely voters. His disapproval rating is roughly equivalent. If we assume that undecided voters who approve of the president will vote for him, and those who disapprove of him will vote for Mr Biden, then Mr Biden, not Mr Trump, stands to gain. An average of YouGov’s polls over the last month yields an eight-point margin for Mr Biden in the standard match-up, but a nine-point margin when undecided voters are allocated according to their feelings toward the president (see chart).
Despite all the hullabaloo about Mr Trump’s allegedly hidden supporters, in the end it may be Mr Biden who turns out to have more than expected.
Editor’s note: This article first appeared in our weekly “Checks and Balance” newsletter on American politics. You can sign up to receive it here.
Read the best of our 2020 campaign coverage and explore our election forecasts, then sign up for Checks and Balance, our weekly newsletter and podcast on American politics.