Dr. Harry Wilson, Political Science Professor at Roanoke College released a poll today which is a second chance to gauge the Virginia Senate race between George Allen and Tim Kaine.
The last poll showed George Allen leading Tim Kaine by 8 points. And as a Conservative, the results of the poll were to my liking politically speaking, I found the methods behind the poll and the demographics to be the first honest poll out in this race.
In that post I wrote:
As always, I look at the numbers behind the polls to see if they pass my “believability” test, or if the results are skewed and need to be “normalized”.
This poll passes with only 1 bit of criticism, sample size.
I am not a fan of polls involving less than 1,000 respondents and this poll interviewed only 607. But to the credit of the Roanoke College, they did a few things that I believe more than make up for a smaller sample size.
One was including cell phones in the sample. 24% of those called were on cells. And they also rotated the order of the candidate names.
And the mix of Republicans (30%), Democrats (28%) and Independents (31%) was pretty good.
My conclusion is that Virginians are beginning to move away from Kaine and towards Allen. In a September 2011 poll, Allen led Kaine 42% to 39% which shows movement away from Kaine and towards Allen.
That poll used a sample demographic of 30% Republicans, 28% Democrats and 31% Independents that looking at the trend in makeup of the Virginia General Assembly as well as the top 3 spots in the state as moving solidly Republican, I believe these percentages accurately reflected election reality of the last 3 election cycles in Virginia.
And going back to the September 2011 poll, the demographic sample size for political affiliation was 29% Democrats and 32% each for Republicans and Independents. And that poll showed George Allen leading Tim Kaine 42% to 39%.
These polls were as “apples to apples” as it gets. The skew in the September 2011 poll was +3 Republican and the poll in the February 2012 poll was +2 Republicans.
We can reasonably disagree about the exact percentages here but based on the last 3 election years since Obama took office, you would have a hard time convincing anyone that any valid poll would favor Democrats as far as demographics. And for the sake of agreement, I could even go along with a point or two skew in favor of Democrats, but I don’t believe that would be as accurate.
So when I saw the latest poll by Dr. Harry Wilson at Roanoke College I nearly choked.
There is no doubt that the dialog surrounding polls by the same group over time will center on movement within that group’s poll. The assumption is that the methodology between the polls will be very similar to one another. And the results show valid movement.
Dr. Wilson knows this.
So after two consecutive polls that favored Republicans slightly, what does Wilson release as a follow up marker in the Virginia Senate Race?
A poll that skews towards Democrats by 9 points – Democrats 36%, Republicans 27% which shows Kaine with a 10 point edge!!!
So now the news will report that Tim Kaine has made an 18 point turnaround since the February and as all media do, they will only report that this is a “same poll movement”.
The massive skew will not be mentioned.
This is both deceptive and irresponsible. Dr. Harry Wilson has now joined the ranks of polls that I consider untrustworthy.
Professor – you get an “F” for this poll. And I hope your students aren’t harmed by this kind of dishonest polling. The world needs more pollsters that do not intentionally skew and manipulate polls to achieve a political objective.
As far as I am concerned, Roanoke Polls are rubbish.
I have sent Dr. Wilson an email and will post his response should I receive one.