And, boom!, just like that, Gov. Glenn Youngkin is suddenly more popular.
Or seems to be.
Yesterday I looked at the results of the recent Wason Center for Civic Leadership poll that found Youngkin’s approval rating to be 41% with his disapproval rating at 43% – meaning just a month into his governorship he was already “under water,” as the political lingo goes.
When I compared this to the approval ratings of other governors across the country, whoa, Youngkin came in last – just below Oregon Gov. Kate Brown, whose approval rating was at 43%. Ergo, my question: Is Youngkin really the least popular governor in the country?
Not long after that column came out, Roanoke College released its own poll on Youngkin’s administration. It has quite different results, ones that Youngkin will like a lot more. The Roanoke College poll puts Youngkin’s approval rating at 50% with a disapproval rating of 41%.
That’s obviously a lot better for Youngkin. Not great, but a lot better – and more in line with last November’s election results where Youngkin took 50.58% of the vote to Terry McAuliffe’s 48.64%. The Roanoke College poll would seem to suggest, at least superficially, that Youngkin’s voters are probably pretty happy with what he’s doing while some of McAuliffe’s voters are still withholding judgment.
So what accounts for the difference between these two polls? Which one is right and which one is wrong?
It’s hard to say any poll is “right” or “wrong.” Sometimes we can, sure. But I’m inclined to think both of these polls are right. How can that be? Is this a Schrodinger’s cat situation? (Don’t know Schrodinger and his famous cat? Here, go have some fun with quantum mechanics.)
The answer lies in who gets polled. In election polls, we often see differences based on the different assumptions that different pollsters have to make. Will this group vote at higher or lower levels than they usually do? That often accounts for differences in polling outcomes that have nothing to do with polls being “right” or “wrong.” In an election poll, they can only be “right” or “wrong” once we see who really does show up to vote. For instance, any pollster who last year assumed that rural Virginians would vote at their normal levels – a perfectly valid assumption – turned out to be wrong because rural voters wound up voting at higher than normal levels. And because the rural vote was overwhelmingly a Youngkin vote, that meant those polls underestimated Youngkin’s strength. That doesn’t mean the polls were biased ideologically, just that reality sometimes turns out different than how we expect. If, four years from now, pollsters use the 2021 numbers as a baseline, they might turn out to be wrong if the rural vote falls back to normal levels. There are ways that pollsters try to gauge the intensity of a respondent’s support and take that into account, but life is often messy.
For what it’s worth, both the final polls that the Wason Center at Christopher Newport University and Roanoke College conducted before the election showed the same thing: Both gave McAuliffe a 1% lead, which is well within the margin of error – listed as plus or minus 3.5% in the Wason Center poll and plus or minus 4.1% in the Roanoke College poll. Given that Youngkin won by 1.94%, it’s fair to say that both polls were, statistically speaking, right. It was going to be a close election and was.
So now let’s turn our attention to these current polls – which aren’t election polls so we have no “reality” to measure them against. Polls are, as we’re often told, “snapshots in time” and opinions often change. Both these polls were conducted at about the same time (Jan. 26-Feb. 15 for Wason, Feb. 7-16 for Roanoke College). Generally a shorter polling window is considered preferable; maybe Youngkin did things during the wider Wason window that changed people’s minds. Also, the Wason polling started just 18 days after Youngkin took office, so Roanoke College gives respondents a bit more time to consider his governorship.
That’s not the important difference, though. This is: The Wason poll found an electorate distinctly more liberal than the Roanoke College poll did.
Of the Wason Center poll’s respondents, 35% described themselves as liberal or very liberal, 24% said they were conservative or very conservative.
In the Roanoke College Poll, 23% of those surveyed said they were liberal, 32% described themselves as conservative.
That right there explains the different outcomes. It stands to reason that a poll with a more liberal pool would produce a less favorable opinion of Youngkin while one with a more conservative pool would produce a more favorable opinion.
So does this mean the Wason Center poll is biased? No. But it does mean the two polls found very different universes. Pollsters can control for certain known variables. We know, objectively speaking, how the Virginia electorate is distributed by age, gender, race, income, those sorts of things. It’s harder to control for ideology because people’s opinions change over time.
For instance, the final Wason Center poll before the election found Virginians to be 23% liberal or very liberal, 40% conservative or very conservative – this month’s poll shows a huge swing, with liberals jumping from 23% to 35% and conservatives shrinking from 40% to 24%. Is that realistic?
Before you say no, consider the Roanoke College poll. Its final pre-election poll found Virginians to be 18% liberal and 39% conservative, so it also found a swing, just not as big – with liberals growing from 18% to 23% and conservatives shrinking from 39% to 32%. Those numbers aren’t nearly as big as the Wason Center poll but they do show the same directional trend. This seems a political application of Newton’s Third Law of Motion: For every action there is an equal but opposite reaction.
There’s also this important caveat: Those pre-election polls attempted to measure “likely” voters, a distinct subset of voters overall, so these polls probably pick up a lot of people who aren’t that inclined to vote. That, too, would explain the higher percentage of liberals in the Wason Center poll because it’s clear that a lot of left-leaning voters stayed home in November. Voters have elected a Republican governor, partly because a lot of Democrats were apathetic about their candidate. Now that they’re seeing him in action, those Democrats aren’t quite so apathetic anymore. (We see the same thing at the national level, just in reverse: It’s why midterm elections typically go against the president’s party, and 2022 looks like it will be particularly bad for Democrats.) The question is not so much which poll is right or wrong on Youngkin’s approval rating, but which poll comes closest to capturing whatever swings of sentiment are taking place?
In any case, let’s take the more favorable (to Youngkin) Roanoke College results and measure them against the approval ratings of other governors like we did yesterday. At 50%, Youngkin is no longer the least popular governor in the country – but he’s still closer to the bottom than he is to the top. By those numbers, there would be nine other governors less popular than Youngkin. That’s still not great.
To be fair, there are a lot of governors bunched together in the 50% to 52% approval rating range – 10 to be precise. The margin of error on the Roanoke College poll is 4.8%, so let’s be generous and say that maybe Youngkin’s approval rating could be as high as 55%. Now there are 18 governors in the 50% to 55% range, so that makes Youngkin’s numbers in the Roanoke College poll look a lot better – better as in pretty average.
It’s also in line with how Ralph Northam polled. At his low point (and we all know what that low point was), he had 32% approval, 39% disapproval. But then he rebounded and at one point hit 59% approval, 29% disapproval. By the time he left office he was at 52% approval, 38% disapproval, so Youngkin has essentially inherited Northam’s numbers (if not the people behind those numbers). If you’re a Republican strategist, you might be more focused on Youngkin’s disapproval numbers than his approval numbers: Northam, at his lowest points, only hit 41% disapproval twice. Youngkin starts there. Maybe Youngkin can only be so popular, but it’s in his interest (and Republicans’ interest) to make sure he’s not polarizing. I’m not sure what the official tipping point for being polarizing is but if I were a Republican, I’d be a little concerned about the disapproval rate. Approval rates can go up with good performance but once a big chunk of people decide they don’t like you, it’s often hard to change their minds.
That’s why the question I posed yesterday is still quite applicable: Is Youngkin popular enough to help Republicans retain the House of Delegates next year and win back the state Senate? Of course, that begs the question: Just how popular does he need to be to accomplish that? Or, conversely, how low does his disapproval need to be?
That we don’t know – and won’t know until next year. Of course, Youngkin’s approval ratings, whatever they may be, don’t exist in a vacuum. There is also public sentiment about President Joe Biden, which right now is pretty bad. Biden’s approval rating in Virginia is 41%, his disapproval rating is 53%. Interestingly, the Wason Center poll has almost exactly the same numbers – 40% approval, 53% disapproval. So if you think the Wason Center poll is off on Youngkin because it’s universe is too liberal, why does it show same numbers as the Roanoke College poll on Biden?
For more context: Biden’s approval/disapproval ratings in Virginia are now what then-President Donald Trump’s were in October 2020 when he was at 41% approval and 56% disapproval in Virginia. The only difference is that those numbers marked Trump’s high point in Virginia during his four years in office while Biden’s mark a low point. Still, we know what happened to Trump – he got voted out. Democrats, take note. Plenty could change in the coming months and years but right now things aren’t looking good for Democrats.
Biden’s numbers are particularly bad when you consider he carried Virginia in 2020 with 54% of the vote, the highest share for any Democratic presidential candidate in the state since Franklin Roosevelt took 57% in 1944.
Now, for the big picture: Youngkin and Biden may not agree on much but they could both commiserate on this point: Voters just seem perpetually unhappy. The Roanoke College poll finds that only 29% of Virginians think the country is on the right track; 67% think it’s on the wrong track. But changing presidents hasn’t really changed those numbers much. Maybe different people think the country is on a different track but when Trump was president, the numbers were about the same or worse. At his best, only 37% of Virginians thought the country was on the right track under Trump; 56% thought it was on the wrong track. At his worst, only 16% thought it was on the right track, with 79% thinking it was on the wrong track.
Same for changing governors, and even changing the party that governor comes from. Virginians started off feeling pretty good under Ralph Northam – 58% right track, 30% wrong track. Then things went south. For a time, the figures fell to 45% right track, 49% wrong track, one of the few times in the history of the Roanoke College poll that a plurality felt Virginia was on the wrong track. Northam turned things around and by last September, the last time the Roanoke College poll asked that question, we were back to 52% right track, 43% wrong track. Nonetheless, Virginians decided to change parties, anyway. Now, in the first Roanoke College poll of Youngkin’s tenure, even though his approval ratings are on the plus side, the right track advantage has disappeared – it’s now 47% each way. It’s tempting to say we’re at a fork in the road.
Whatever the cause of voters’ discontent (and there are surely many, from across the ideological spectrum), it seems fair to ask: Just what is it that would make people happy? Because changing presidents and changing governors and changing parties sure doesn’t seem to be it.