Donald Trump has shocked the world and won the presidential election. But you shouldn’t blame most of the polls for your surprise. Polling errors of about three percentage points are fairly common, as we’ve pointed out at THE WEEKLY STANDARD and as Nate Silver has explained at FiveThirtyEight.com.
And most of the polling errors in the 2016 presidential election were within three percentage points.
The Real Clear Politics average of national polls showed Trump down by 3.3 percentage points; the New York Times still projects Trump will lose the popular vote by 0.6 points. And most of the battleground states that decided the winner of the Electoral College were pretty close to the final RCP polling average (with one big exception).
Florida: The RCP polling average showed Trump leading by 0.2; he’s projected to win by 1.4.
New Hampshire: The RCP polling average showed Trump down by 0.6; he’s projected to lose by 0.1.
North Carolina: The RCP polling average showed Trump up 1.0; he’s projected to win by 3.8.
Nevada: The RCP polling average showed Trump up 0.8; he’s projected to lose by 2.4.
Colorado: The RCP polling average showed Trump down 2.9; he’s projected to lose by 3.9.
Pennsylvania: The RCP polling average showed Trump down 1.9; he’s projected to win by 1.0.
Michigan: The RCP polling average showed Trump down by 3.4; he’s projected to win by 0.2.
Ohio: The RCP polling average showed Trump up by 3.5; he’s projected to win by 8.5.
Wisconsin: The RCP polling average showed Trump down by 6.5; he’s projected to win by 0.9.
As you can see, Trump under-performed his polling averages in Nevada and Colorado, and he exceeded expectations the critical states he needed to win the Electoral College. Michigan is the perfect example. And the only huge polling miss in the presidential race was Wisconsin, where even the gold-standard Marquette poll finally got it wrong.
