Donald Trump has been advancing two arguments about the polls that increasingly suggest he will go down to defeat. The first, and best-covered, is that the election will be “rigged” and that the polls may well be massaged to enable that. The second is that the polls that have shown him losing are not much different from the ones that traced his rampage through the Republican primaries. At many of his rallies, Trump asks crowds to remember that polls showed him losing states and that the media dug his grave, yet he went on to win.
One problem: This is mostly not true. Although Trump clinched the Republican nomination with a larger-than-expected victory in Indiana's May primary, public polling usually got his numbers right. It more frequently overstated his support than understated it.
In Iowa, the polling average found Trump in the lead by 4.7 points — representing a small bounce from what had seemed to be the peak of Sen. Ted Cruz (Tex.) two weeks earlier. Cruz ended up winning by 3.3 points, a result that stirred some false hope that Trump's support would collapse across the rest of the early-primary states.
It didn't. In New Hampshire, Trump led by an average of 17.2 points; he won by 19.5 points. In South Carolina, Trump led by 13 points and won by 10. Nevada was too lightly polled for RealClearPolitics to create an average, but his 22-point win there was slightly smaller than the final public poll — by CNN and ORC — projected.
The results on March 1, Super Tuesday, were all over the map, and a few states such as Vermont and Alaska were barely polled at all. But across every Super Tuesday state, Trump's average polling lead was 9.7 points. On Election Day, he beat expectations in several Deep South states, but he underperformed them in Alaska, Oklahoma, Texas, Vermont and Virginia for an average win of 5.6 points. He underperformed by a little less in the March 8 contests, the first with a shrunken four-way field. And on March 15, the set of contests that drove Sen. Marco Rubio (Fla.) from the race, Trump basically ran along with the polling in Florida and Illinois but was far behind it in Missouri, Ohio and North Carolina.
At the time, Trump's rivals — who invested more in traditional polling — basically got his margins right. One reason that Cruz spent some of the run-up to the March 15 contests in Florida and Illinois was that his polling showed what the sparse public polls did not, a surge in his direction in Missouri and North Carolina. When Cruz narrowly lost both states, the theory was that voters moved Trump's way in a backlash to the highly publicized protests that violently disrupted a Trump rally in Chicago. When Cruz was on track to losing Indiana, private polling showed what public polls did not — the bottom had fallen out, and Trump was going to win big.
Plenty of campaigns close out the dark, waning days by saying that polls will underestimate their support. Sometimes, those campaigns are proved right. But the copious evidence of the long Republican primary season suggests that Trump's support in polls roughly matches his total vote.