I grew up in an era where the principal at McGaheysville Elementary School in Rockingham County kept a paddle prominently displayed on the wall in his office.
I was never on the receiving end of this instrument, but the mere presence of the paddle was known to strike fear into many of us schoolchildren. I don’t remember anyone in my class suffering such punishment, but now and then we’d hear a rumor of some unfortunate lad being summoned to the office. Assume the position. Those were different days.
I don’t have a paddle. Those are frowned upon nowadays. These days, I don’t even need a pen, but I do occasionally feel compelled to use the keyboard to take a proverbial paddle to those who dare to insult our part of Virginia.
When a Richmond journalist dismissively wrote about “whatever the hell is west of Roanoke,” I educated her.
When Hung Cao, the Republican candidate for the U.S. Senate, called the Staunton newspaper “podunk” and said Abingdon was too far to drive for a campaign forum, I called him to task. I did so again when he called everything south of Loudoun and Fairfax counties “southern Virginia.”
Today a Democrat — former Gov. Terry McAuliffe — is getting called out.
The recent Roanoke College poll showed Kamala Harris leading Donald Trump by 3 percentage points in Virginia. This was better than the tie that Joe Biden had, but it’s still within the margin of error. I know many people on both sides were surprised that Harris didn’t have a wider lead.
One of those was McAuliffe, who proceeded to criticize the poll: “That was a Roanoke College poll, who has never had a poll that’s been correct in their life,” he told The Washington Times. “When I ran for governor, they had me losing by 500 points, and I won the state.”
Now, it’s certainly fair to criticize polls, and I realize the former governor is employing some hyperbole here, but I feel compelled to set the record straight.
First, let’s look at McAuliffe’s claim that Roanoke College “never had a poll that’s been correct in their life.” This is personal to me: Back in 1994 when Virginia had a very contentious U.S. Senate race, the Roanoke College poll was the first to report that Democrat Charles Robb had pulled ahead of Republican Oliver North. I wrote a story for The Roanoke Times that quoted numerous experts criticizing the poll as an outlier — until then, North had consistently led. Turns out, the Roanoke College poll was right; it was simply the first to catch a late-breaking wave toward Robb. Lesson learned.
Just for kicks, though, let’s look at the poll’s record:
2021 governor’s race
The final Roanoke College poll in the governor’s race had McAuliffe leading Republican Glenn Youngkin 47% to 46% — definitely within the margin of error. That was consistent with other polling at the time, which had McAuliffe slightly ahead until the final days when Youngkin pulled out in front — and won 50.6% to 48.6%. According to the data website FiveThirtyEight, all the other polls conducted in the same timeframe also showed McAuliffe up by 1 percentage point except one. That was a Fox News poll that showed Youngkin up by 8 points. After the Roanoke College poll, four of the next five polls showed Youngkin narrowly ahead. The Roanoke College poll was fundamentally right — the race was close. It just came out before that final Youngkin surge. (This also highlights a key point that must be made about every poll: They are not necessarily predictive tools; they are snapshots in time. Campaigns are dynamic events. Things change. That’s the whole point of campaigns: To change people’s minds.)
2020 presidential race
In the final poll before the presidential election, the Roanoke College poll showed Biden leading Trump in Virginia 53% to 42%. The final result was 54.1% to 44%.
2020 U.S. Senate race
In the final poll, the Roanoke College poll showed Democrat Mark Warner at 55%, and Republican Daniel Gade at 39%. The final result was Warner 56% and Gade 43.9%.
2017 Governor
In the final poll, Roanoke College found the race between Ralph Northam and Ed Gillespie tied. That was consistent with other polling at the time. Ten polls taken about the same time showed Northam with an average lead of 3 percentage points, within the margin of error. Two polls showed Gillespie ahead. Northam eventually won 53.9% to 45.0%, which suggests that late-breaking votes went Northam’s way.
2013 Governor
As for McAuliffe’s claim that the Roanoke College poll had him losing to Republican Ken Cuccinelli in 2013, the best I can say is that McAuliffe is mistaken. The first Roanoke College poll in that year’s race did show Cuccinelli ahead, but subsequent ones showed McAuliffe pulling ahead.
July: Cuccinelli 37, McAuliffe 31
September: McAuliffe 35, Cuccinelli 33
Early October: McAuliffe 40, Cuccinelli 34
Late October: McAuliffe 46, Cuccinelli 31
The most egregious claim in McAuliffe’s poll critique is this: “Mr. McAuliffe said Roanoke College’s methodology was flawed because 51% of the poll was conducted outside heavily Democratic regions in the state, including Hampton Roads, Northern Virginia, and Richmond.”
This is simply untrue.
Roanoke College publishes the details about its poll. Those show that only 29% of the poll respondents are “outside heavily Democratic regions in the state, including Hampton Roads, Northern Virginia and Richmond.”
The precise breakdowns:
Northern Virginia 32%
Tidewater 23%
Central Virginia/Richmond 16%
Southwest Virginia 12%
Southside 9%
Shenandoah Valley 8%
Some polls “weight” respondents to match known geography — in other words, if there aren’t enough answers from a certain part of the state, the poll gives those answers more weight so the geography works out. The Roanoke College poll does not do that, says poll director Harry Wilson. It assigns a quota system and polls a specific number of people in each geography. “Simply put, the methodology about the region is the same as we used back when Mr. McAuliffe was governor, running for governor (both times), and before that,” Wilson said.
McAuliffe may not like the results because they don’t show Harris up by as much as he would like, but his critique of the poll’s methodology is simply wrong.
Want more political news and analysis? I write a weekly political newsletter, West of the Capital, that goes out every Friday afternoon. You can sign up for that or any of our other free newsletters here:
