How Did the Polls Do in Virginia?

The election results in Virginia aren’t official yet, but they’re very close to final and unlikely to change much. So…while we’re sitting around waiting for the ridiculous TV networks and AP to proclaim Joe Biden the next president – something that has been obvious for hours now – let’s take a look at how the polls did in Virginia  to see how they did.

First, the results:

• President: Joe Biden 53.77%-Donald Trump 44.37% (Biden +9.4 points)
• U.S. Senate: Sen. Mark Warner 55.6%-Daniel Gade 44.3% (Biden +11.3 points)
• VA01: Rep. Rob Wittman 58.4%-Qasim Rashid 41.4% (Wittman +17.0 points)
• VA02: Rep. Elaine Luria 50.9%-Scott Taylor 46.5% (Luria +4.4 points)
• VA05: Bob Good 52.6%-Dr. Cameron Webb 47.2% (Good +5.4 points)
• VA07: Rep. Abigail Spanberger 50.5%-Nick Freitas 49.4% (Spanberger +1.1 points)
• VA10: Rep. Jennifer Wexton 56.4%-Aliscia Andrews 43.5% (+Wexton +12.9 points)

Now, how did the polls do? See below for the Virginia polls from September until November 2, courtesy of FiveThirtyEight.com.

• President: The final polls by Swayable (Biden +15 points) and SurveyMonkey (Biden +16 points/+17 points) were waaaayyyy too bullish on Biden, who appears to have ended up winning by 9.4 points. Data for Progress and Roanoke College were closer, with Biden +11 final polls. VCU’s and CNU’s final polls had Biden +12 points, which is 2.6 points too high…not too bad. And the WaPo/GMU’s final poll had Biden +11 points, which is just 1.6 points too high. Actually Cygnal nailed it, with their final polls on October 19 showing Biden +9 or +10, which is exactly right. As for Civiqs, they were too optimistic, with Biden +13 in their final poll. So the winner here appears to be Cygnal, followed by WaPo/GMU. The loser is SurveyMonkey, followed by Swayable and Civiqs.
• U.S. Senate: Swayable’s final poll had Sen. Warner winning by 21 points, which is 9.7 points too high. Yikes. Data for Progress was better, with Warner +15 points, just 3.7 points too high. Roanoke College had Warner +16 points, which is 4.7 points too high. VCU had Warner +17 points, which is 5.7 points too high. CNU had Warner +20 points, which is 8.7 points too high. The WaPo/GMU had Warner +18 points, which is 6.7 points too high,. Cygnal did well again, with Warner +7 or +9 points, which is just a bit low, actually. Nice job! And Civiqs did pretty well here, with Warner +11 points, just 1.3 points too high.
• VA01: There were no public polls in this race, although I hear an “internal” had Rashid down just 7 points. That one was too optimistic by about 10 points.
• VA02: There was only one public poll here in the fall, and that was CNU, which had Rep. Luria up 7 points. That’s 2.6 points too high, which isn’t horrible.
• VA05: Public Policy Polling’s final poll of this race had Cameron Webb +3 points, which got both the winner wrong and also was off by a whopping 8.4 points. Really bad. Other than that, Global Strategy Group did three polls of VA05 in the fall, with the latest one showing Webb +2 points, the one before that showing Webb +3 points, and the one before that with Good +1 point. In other words, all of these were off badly, with only one (the Global Strategy Group poll in mid-September) showing Good winning. #FAIL
• VA07: There were no public polls of this race, which kind of mind boggling, actually, given how important and high-profile it was.
• VA10:  There was only one public poll here, by Garin-Hart-Yang Research Group, and it showed Wexton up 22 points, which turned out to be about 12 points too high. At least they got the winner right, though! LOL