The Trafalgar Group’s final VA Governor poll shows Youngkin up 2.3 over McAuliffe. Although I think the poll is accurate, there are sampling problems in the poll that make me cautiously optimistic about the results.
One polling company stood out as being one of most accurate and interesting towards the close of 2016’s Electoral Cycle. It was The Trafalgar Group. Many mainstream media people laughed at the little-known firm and mocked it as a right-wing joke. At the end of Election Night 2016, the Trafalgar Group came out as the most precise firm in 2016.
While I have questioned the methodology and results of some polls they have done since then, much to the friendly chagrin of Robert Cahaly, Trafalgar’s Chief Pollster, I have consistently found Trafalgar’s work among the most accurate, and certainly the most “accessible” in the terms of talking to the people behind the polls for clarity.
In this most recent and final Virginia Governor’s poll, Trafalgar shows Glenn Youngkin, the Republican, leading Terry McAuliffe, the Democrat, by 2.3 points. With the election held tomorrow, Youngkin’s lead should indicate a significant chance of victory, especially considering everything that Trafalgar got right with this poll.
Like I said, Trafalgar is my top choice for accuracy in polling since 2016. This is largely due to the fact that their sampling and methodology yields the best results. This trend is evident in this latest poll. The poll’s sampling, which covers everything but age, is perfect. A Gallup poll in 2017 found that Virginia’s party split was 45 to 38 favoring Democrats. Virginia does not have a party register voter. The poll is almost identical. Both the demographics of gender and ethnic heritage are very accurate. Furthermore, the sampling from Virginia’s 11 Congressional Districts seems relatively well split, not favoring any one area over another. Shifting priorities can result in geographic sampling errors that could lead to inaccuracies. Urban Democrats might be more liberal than rural Democrats, but they may also be more conservative. If you do not oversample a particular geographic region, this can result in those errors. Most polls assume party affiliation to be the final factor that influences ideological priorities. One could have a lot confidence in the accuracy and precision of the methodology and sampling.
However, even though we strive to achieve near perfection, it can sometimes be difficult to tick all the demographic boxes. It is also more challenging to locate specific voters who meet the requirements of the poll. This can lead to good factors such as the ones you have seen above and oversampling other factors which may lead to inaccuracies. There are two problems I see in this poll. Neither is sufficient to discredit the poll. They could even cancel each other.
First, the poll found that 41.3 percent were aged between 45 and 61. However, only 23 percent Virginians are from this age. A further 25.8% of respondents were aged 65 and over, even though only 11.2 percent of Virginia’s population is from this age group. It is clear that older voters are more likely to vote, which means it’s a good idea to lean in this direction. Is this a problem? If it were just that older voters vote more often it would work, it might be fine. But older voters are also more conservative and this creates an ideological bias. This poll may show a bias such as this. It could indicate that there is a much larger race than the 2.3 percent, or even a tie or slight Dem advantage.
But! Cahaly, don’t panic! Don’t kill me yet!)
From October 29th to the 31st (a Friday through Saturday), this poll was taken. Is that really significant? Two things are apparent in this poll. These dates could explain why Trafalgar couldn’t contact younger voters. The majority of younger voters have children, so they were likely not at home on a Saturday or Sunday. This could also explain why Trafalgar was unable to contact more young voters. This is why including those voters may not have been possible. While that doesn’t completely erase my concerns about the older voter oversampling, there is an issue that will: The Ole October surprise, and this one happened on the 29th.
As you may have read here at RedState (Our Bonchie covered it here) or elsewhere, a shocking sight of a group of White Supremacists at a Youngkin Rally, led many to believe that the Republican’s chances for a victory were sliding away. That was until, those same white supremacists, including a Clayton-Bigsby-looking fellow (bonus points if you get the reference), were all identified as VA Democrat staffers. After chaos ensued, and denials of involvement from the campaign and the party, the losers at the Lincoln Project came out to claim credit for the alleged “demonstration.” While I and others still question how the Lincoln Project had access to staffers from both the campaign and the party (a potential campaign finance violation), the damage still fell hard on McAuliffe’s campaign. This failed stunt may have led to a serious slide towards Youngkin. The polling could have taken place before some respondents responded. Some polling did not include younger voters, who may have had their opinions affected by the error.
If anything, any potential issue with the poll not accounting for some more liberal voters (due to age sampling) is washed by the Democrat’s failed white supremacist stunt. This may also eliminate my concerns about Youngkin’s age-sampling to gain an advantage. We will see.
As a result of the poll’s fantastic methodology and with my two concerns washing out, I give this poll a solid A grade.
Now, if only Cahaly’s Braves can pull off a win in the World Series, he might just have a very good week! (And that’s coming from a very sore Dodger fan!)
About Post Author
