So How Did the Pollsters Do in the 2024 Elections in Virginia?

Now that we have close-to-final election results (see VPAP for the exact numbers), we can evaluate how the pollsters faired.

Here in Virginia, polls released in October or early November (source: FiveThirtyEight.com) had the race for president as follows. Note that Kamala Harris won Virginia by approximately 5.6 points.

• AtlasIntel (11/1-11/4): Harris +6
• Research Co. (11/2-11/3): Harris +6
• Chism Strategies (10/28-10/30): Even
• Cygnal (10/27-10/29): Harris +7
• Roanoke College (10/25-10/29): Harris +10
• ActiVote (10/2-10/28): Harris +8
• CCES/YouGov (10/1-10/25): Harris +8
• Quantus Insights (10/22-10/24): Harris +1
• WaPo/GMU Schar School (10/19-10/23): Harris +6
• CNU Wason Center (9/28-10/4): Harris +11

So, given that Harris won by nearly 6 points (~5.6 points at the moment), it looks like AtlasIntel, Research Co. and the WaPo/GMU Schar School nailed it, with Cygnal off by just 1 point, ActiVote and CCES/YouGov off by just 2 points. Not bad! In contrast, Chism Strategies missed by 6 points (too low for Harris), Quantus Insights missed by 5 points (too low for Harris), CNU’s Wason Center missed by 5 points (too high for Harris), and Roanoke College missed by 4 points (too high for Harris).

As for the U.S. Senate contest between Tim Kaine and Hung Cao, polls released in October or early November had it as follows. Note that Kaine won by nearly 9 points (8.6 points at the moment).

• AtlasIntel (11/1-11/4): Kaine +9
• Research Co. (11/2-11/3): Kaine +12
• Chism Strategies (10/28-10/30): Kaine +1
• Cygnal (10/27-10/29): Kaine +10
• Roanoke College (10/25-10/29): Kaine +11
• ActiVote (10/2-10/28): Kaine +9 (an earlier poll by ActiVote had Kaine +12)
• WaPo/GMU Schar School (10/19-10/23): Kaine +13 or Kaine +14
• CNU Wason Center (9/28-10/4): Kaine +20

So for U.S. Senate, given that Kaine won by 8.6 points, it looks like the AtlasIntel and ActiVote basically nailed it, with Cygnal off by just 1.4 points (too high for Kaine), Roanoke College off by 2.4 points (too high for Kaine), and Research Co. off by 3.4 points (too high for Kaine). Pollsters not doing as well were Chism Strategies (nearly 8 points too low for Kaine), CNU’s Wason Center (over 11 points too high for Kaine) and WaPo/GMU Schar School (about 5 points too high for Kaine).

Finally, regarding U.S. House races, there wasn’t much public polling. For VA07, the only ones listed by FiveThirtyEight in October were Ragnar Research Partners, which did “internal” polls for Republican Derrick Anderson’s campaign. And actually, they did pretty well, with their final poll showing Eugene Vindman up 2 points – very close to Vindman’s winning margin of 2.5 points (Ragnar’s previous poll had Vindman up 1 point, which is still not too shabby). As for the other top-tier race in Virginia, VA02 (between incumbent Rep. Jen Kiggans and Democratic nominee Missy Cotter Smasal), an “internal” DCCC poll (10/19-10/20) had the race tied, while CNU’s Wason Center poll (10/11-10/20) had Kiggans up 1 point (Wason’s previous poll, in September, had Kiggans up 5 points). In the end, Kiggans won by just over 4 points, so Wason’s September poll (Kiggans +5) was very close, with its October poll three points too low for Kiggans. As for the DCCC poll, that was too optimistic (by ~4 points) for Smasal.

P.S. Of course, it’s possible to be right for the right reasons, right for the wrong reasons, wrong for the right reasons or wrong for the wrong reasons. And then there’s also just plain luck – or lack thereof.