The Washington Post pays for polls.
The polls aren’t worth much if they can’t get a story out of them.
So, how do they get a story? They fudge the numbers and spin them. Then, they release them bit by bit, like the drip…drip…drip of Chinese water torture to demoralize the opposition. And they hope no one looks behind the curtain because the old man pulling the levers doesn’t want his bias know.
We’re looking behind the curtain.
It started last week when the WaPo rolled out the Presidential numbers for Virginia showing Obama up on Romney by 7%, 51-44 among Registered voters. Fault one – not LIKELY voters, Registered, ignoring any enthusiasm gap.
With any poll you have to look past the top line and into the crosstabs to see how these numbers were created. How many men vs. women, what was the race split, what portion of those polled came from where in the state. All very important numbers that the Washington Post has not yet released. Here’s the headline, spun the way they’d like it, but the facts will have to wait for another day.
But what they did release is damning enough. And not for Romney.
Democrat – 31%
Republican – 25%
Independent – 36%
The last time Republicans accounted for only 25% of the electorate LBJ was elected President in 1964. In 2008 Republicans accounted for 33% of the electorate. Out the gate Republicans were under-polled by 8% compared to 2008, by more if you look at turnout for 2009-2011.
Now, as bad as 25% is, it’s a modest improvement over the May 2011 poll WaPo ran that only surveyed 22% Republicans.
Democrat – 31%
Republican – 22%
Independent – 36%
Last year’s survey was the starting point for today’s Washington Post article that said Virginia’s were less optimistic about Gov. Bob McDonnell and the state’s direction. Note, it’s not that they are NEGATIVE about McDonnell or the state’s direction, but that they are not as optimistic as they were a year ago.
McDonnell’s approval rating stands at 56%, a 6 point drop from a year ago, while his disapproval is at 35%. He has a net approval of 21%. Down, yes, but still leaps and bounds above the President’s 53% to 44% for a net +9% approval which is static to a year ago.
Virginian’s approve of Gov. McDonnell’s job performance 26% more than they do the President of the United States’ despite polling 7% MORE Democrats than Republicans, who were undersampled compared to the true electorate by as much as 14 points. Yet the Washington Post wants to spin this into a negative for the Governor.
Want more reasons why the Washington Post poll is terrible?
May 2011 Support for Obamacare (+9% Dem):
NET Support – 42%
NET Oppose – 48%
NET -6% opposition
May 2012 Support for Obamacare (+7% Dem):
NET Support – 49%
NET Oppose – 45%
NET +4% support
The Washington Post wants us to believe that in the last twelve months support for Obamacare has swung a full 10% into its favor. Really.
Oh, and how are these numbers measuring up to other polls? For this exercise we’ll include Senate numbers because you know they’re coming from the WaPo.
April 2012: Rasmussen (Likely Voters)
Obama – 44%
Romney – 45%
Allen – 46%
Kaine – 45%
Obamacare (question was should it be overturned):
Overturn – 49%
Uphold – 39%
Democrat – 33%
Republican – 33%
Independent – 34%
May 2012: Public Policy Polling (Registered Voters)
Obama – 48%
Romney – 42%
Allen – 46%
Kaine – 45%
Democrat – 39%
Republican – 32%
Independent – 29%
Real Clear Politics (averaging polls):
Obama – 47.5%
Romney – 44.3%
Allen – 43.7%
Kaine – 44.2%
The Washington Post is setting itself up to be an outlier that resembles polling from biased firms. Say what you will about Public Policy Polling’s bias (they are an unabashed Democratic polling firm that works with organizations like DailyKos and SEIU) but at least their numbers somewhat resemble 2008’s exit polls.
The Washington Post, on the other hand, has surveyed a fantasyland that only represents Virginia if it was populated solely by their own editorial board. These numbers don’t serve to provide a service to Virginians but to create a story out of thin air of a Democratic dominance that just does not exist within the Commonwealth. The Washington Post wanted a story that could sell and their pollster delivered numbers that only serve to try and demoralize Republicans and energize Democrats.
But those numbers just don’t match up with reality. And the Washington Post is hoping that no one pulls back the curtain to see the truth.
UPDATE: The Hill points out that Gov. McDonnell’s average over the last four polls puts his approval rating at +23.5%, despite “popular spin” that he’s tanking horribly. Even the WaPo can’t spin their way out of this, no matter how they want to stack the numbers against the GOP.
4:33PM UPDATE: RPV released a memo from Pete Snyder, Chairman of Virginia Victory 2012:
To: Interested PartiesFrom: Pete Snyder, Chairman Virginia Victory 2012Subj: Great numbers for Governor McDonnell and Republicans, horrible news for Barack Obama!
Last week the Washington Post conducted a statewide poll in Virginia and has been slowly releasing the results over the past few days. While we are still six months away from Election Day I will guarantee you the horse race numbers of this election will change week by week, if not day by day. Sometimes our folks will be up and sometimes they will be down. That is to be expected in a swing state like Virginia.
However, the results of the recent Washington Post poll show some encouraging signs for our Republican team… but you might not be able to tell that from the newspaper headlines.
When I took a look at the topline numbers and studies the sample universe I was stunned at the methodology the Washington Post’s used for this latest poll. Upon further review, I was elated at the results for our Governor and Republican candidates and have a better understanding of why the Obama/Kaine team is panicking. Please consider:
• The people who responded to the poll weren’t likely voters, and many weren’t even registered voters. According to the Washington Post, the poll was a “random sample of 1,101 adults in the Commonwealth of Virginia.”
• Consider the Exit Polls conducted during the last three statewide elections in Virginia
o 2006 – 39% Republican, 36% Democrats, and 26% Independents
o 2008 – 33% Republicans, 39% Democrats, and 27% Independents
o 2009 – 37% Republicans, 33% Democrats, and 30% Independents
The recent Washington Post poll sample only has 25% of the interviews being completed by Republicans, yes just 25%. – The last time Republicans comprised less than 30% of the general election turnout was more than 40 years ago.
What this poll shows is that the Washington Post has dramatically under sampled Republican voters, severely skewing the results. If you were to take the average Republican turnout in the last three statewide elections (36%), the Washington Post under polled Republicans by more than 11% which is a substantial factor in this flawed poll.
The sample that the Washington Post used for their poll creates an environment for Republicans that is far worse than those in some of the bluest of blue states in the Democratic wave year of 2008 — including Illinois (28%) and California (30%).
Their flawed methodology in more in line with the political realities of Vermont (23%) than the Commonwealth of Virginia.
The fact that Governor McDonnell’s favorable rating remains high despite the flawed sample shows that folks approve of his efforts to rebuild our economy and provide more opportunity for all Virginians.
And the White House should be very concerned by this severely flawed poll. Even with a poll that is unquestionably tilted in their favor, Barack Obama’s is barely able to get a positive approval rating.
Further, the simple fact that this poll includes unregistered voters, to say nothing of the fact that it doesn’t even reflect a realistic 2012 turnout sample, the overall results should be completely ignored.
It’s unfortunate that the Washington Post drew such a bad sample and still chose to run with the results. Random samples are just that, random, but pollsters should check their results against reality before rushing to print.
But if you take away nothing else from this memo, remember this: Even in a skewed sample like this, Governor McDonnell still has a higher approval rating than President Obama.
Why? The Governor’s focus on jobs and the economy and keeping taxes low and getting spending under control is working. We’ve got budget surpluses and falling unemployment. That’s the same focus we see from Governor Romney, and that’s why he’s in such a great position at the start of the general election. That’s the approach voters want to see; it’s what they haven’t seen from the Obama Administration over the past 3 and a half years.
Given that the eyes of the country will be focused squarely on our Commonwealth over the next six months, here’s hoping anyone who is conducting public opinion research in the Commonwealth in the weeks ahead will give their samples a good solid reality check before making it public!