Wonder why nobody believes polls? I don’t wonder. Polling has become a remarkably inaccurate science lately. Skewed methodology will give skewed results, and the Virginia Democratic primary is chock full of them.
As I mentioned in a previous column, I didn’t get too excited by polls showing Northam vs. Perriello tied. Quinnipiac and CNU’s Wason Center showed that in their polls of registered voters, with CNU including a large swath of voters who weren’t interested enough to vote last year.
Public Policy Polling released a poll this week showing LG Ralph Northam with a 14 point lead, in contrast to the others. A few reasons I believe this poll instead of the others:
- It’s a likely voter model. It doesn’t include non-voters. Virtually every respondent voted in the Democratic Presidential Primary, and 2-1 voted for Hillary, which is reality. This poll gives results from a voter pool that is very likely to turn out on June 13th.
- The demographics match Democratic turnout. Voters under 40 make up only 17%, where previous polling nearly doubled the percentage of young turnout. This poll has the age split in a more typical proportion.
- Northam dominates Hampton Roads and Richmond. Perriello dominates Roanoke and Charlottesville. There’s a battle in NOVA, with the two virtually tied. That equates to a Northam statewide primary win, and by a comfortable margin.
- Perriello leads among younger voters, but by half the margin of the previous polls. That’s not movement. That’s just accuracy in this poll compared to polling non-voters in the other polls. Northam leads among all others, and crushes Perriello by 24 points among voters over 60.
- Northam wins the female vote by 20 points, and the male vote is split. Again, given Northam’s image and record, that’s believable and not likely to change.
- Perriello wins Sanders voters 47-29. Northam wins Hillary voters 47-19. Guess which voting block is larger.
With virtually nothing unchanged in the Democratic Primary campaigns, two recent polls with obvious flaws show ties, but this poll shows a lopsided Northam victory. That’s the damage that flawed skewed methodology can do to public confidence in polling, which must be at an all-time low since their performance in 2016.
Perriello and his supporters will claim bias in that the poll was paid for by the Northam-supporting Virginia Education Association, and that Perriello will turn out unlikely voters and turn polls like this on their heads.
VEA bias? Could be, but it’s not evident in the questioning or the demographics.
Will unlikely voters become likely? Possibly. In the Forbes v. Taylor Republican primary, polls focused on traditional primary voters missed large swaths of voters that Taylor turned out in his victory that wasn’t even close.
But Perriello will have to turn out a groundswell to combat Northam’s establishment lead, and despite their liberalism, Virginia Democrats can be remarkably establishment in primaries.
With less than two months to go, Perriello would need either a Northam collapse or a head-shocker reversal in turnout to win.