Is there something funny going on with the surveys Public Policy Polling (PPP) is releasing here in New Mexico?
PPP is a national outfit associated with the Democratic Party. The North Carolina-based firm doesn’t hide that fact but insists it doesn’t allow its rooting interest for Democrats and liberal causes interfere with its polling data.
Republicans often cry foul when PPP numbers show GOP candidates and issues in less than favorable light — that’s natural in the zero-sum game of politics. But conservatives complain that PPP surveys aren’t just snapshots of elections and issues but, rather, an attempt to advance talking points to try to shape opinion and influence the media.
“They’re not so much a polling organization as they are a group of people trying to establish a narrative,” one statewide Republican official who preferred not to have his name used told Capitol Report New Mexico this week.
PPP Director Tom Jensen has said that’s simply whining.
“It’s fine with me if your readers want to not believe our polls,” Jensen told the conservative magazine The National Review in 2010, “but we have no history of cooking the numbers for Democrats.”
But a couple issues have come up in polling that PPP released this week for New Mexico.
For example, on Wednesday (April 25), PPP released some numbers about the Barack Obama-Mitt Romney race.
While releasing the numbers, Jensen declared in a news release that “New Mexico is not going to be a swing state this year,” pointing to its survey showing President Obama with a 14-point lead in the state over Romney.
Jensen also pointed out that, according to PPP’s numbers, Republican Gov. Susana Martinez wouldn’t help Romney in New Mexico should the former Massachusetts governor select Martinez as his voice for Vice President.
“Obama’s approval numbers in the state are on the rise,” Jensen wrote. “53% of voters think he’s doing a good job to 44% who disapprove. That’s up from a 49/46 spread in December. The movement has been mostly due to a trend we’re seeing across the country- Democrats unifying around Obama.”
But just a few days earlier, another PPP survey apparently came out that raises some questions.
We say “apparently” because the numbers have not been released — as far as we can tell — on the PPP website or to the media at large.
Instead, the numbers were posted on the Internet site Scribd.com and show that in a poll of 587 Hispanics in New Mexico, Gov. Martinez has an approval rating of 57 percent — a sky-high number for a Republican:
These numbers come from a phone survey of 587 New Mexico Hispanics taken April 16 and 17.
Is this poll really from PPP? The chart uses the same typeface and fonts as PPP and the firm’s logo, address and phone number are listed at the bottom of each page. (Click here to look at the document for yourself.) Capitol Report New Mexico tried to get confirmation from Jensen but two voicemails left at his number at the firm have gone unreturned.
So Martinez has a 57 approval rate among Hispanics in New Mexico in the April 16-17 poll but in a different poll released Wednesday (April 25), the governor’s approval number with Hispanics is at 51 percent:
These numbers came from a phone survey of 526 New Mexico voters regardless of ethnicity and then were broken down along gender and ethnic lines in the survey’s cross-tabulations. This poll was conducted April 19-22.
What accounts for the six-point difference? And why would it swing from 57 down to 51 between April 16-17 and April 19-22?
Now this is not a potential scandal of Woodward and Bernstein proportions.
After all, we’re talking about two separate polls. And whether the figure for Hispanic approval for Martinez is 51 percent or 57 percent, her numbers are still impressive.
“But there are a couple problems with this thing,” a New Mexico Republican said. “First, why didn’t they highlight the poll that showed her [Gov. Martinez] at 57 percent? That’s pretty noteworthy, don’t you think? And second, what does it say about their pollling that the numbers shift six percentage points in the space of just a couple days?”
Good questions that we would have liked to have asked Jensen if he had returned our calls.
In fairness, in his release on Wednesday, PPP’s Jensen wrote of the survey’s Romney-Martinez numbers:
That still doesn’t take away from how impressive Martinez’s numbers are. Her 54% approval rating ranks her 9th in the country out of more than 40 Governors PPP has polled on in the last couple years and the fact that she’s doing it even as the state prepares to vote overwhelmingly for Obama is all the more striking. She has an outstanding 60/26 spread with independents and a much higher than normal level of crossover support with 31% of Democrats approving of her.
In addition, the Obama lead of 14 points on Romney tracks with the numbers that Rasmussen Reports (a polling outfit that Democrats accuse of skewing in favor of Republicans) released just two weeks ago showing Obama up by 16 points.
Then there are some questions about PPP’s methodology.
Republicans across the country have accused PPP of weighting their numbers too heavily towards Democrats. (Earlier this year, the New Mexico GOP sent out this news release blasting PPP after it came back with some numbers Republicans said undercounted Martinez’ support.)
Haussamen questioned the Democrat/Republican distribution to Jensen, who raised some eyebrows by revealing that “we don’t weight for party at all, for gender and race we weight to a range, depends on the response for that individual poll.”
Haussamen: Can you explain more about what you mean by “weight to a range?”
Jensen: generally 35-40 pct are going to be Hispanic and 51-55 pct are going to be women…but we don’t predetermine an exact number
Haussamen: I thought many polling companies weighted pretty exactly for party registration or turnout. Am I wrong?
Jensen: there are as many weighting strategies as there are polling companies
PPP has defended its methodology by pointing to its successes.
Last year, Politico praised PPP for being “within one point in one special House election, within two points in another and within three points of calling a gubernatorial primary” in races across the country.
But Republicans have accused PPP of propping up storylines early in an election cycle and then retreating to more accurate numbers as balloting draws near.
In Wisconsin, liberals have been calling for the head of Republican Gov. Scott Walker, who has drawn their wrath for instituting changes to public employees pensions, and Democrats have mounted a recall effort.
In previous posts and polls on its website, PPP has noted what it said were weaknesses in Walker’s support.
“Walker would lose rematch with Barrett in Wisconsin” it declared in one post on March 1, 2011.
“Walker would lose to Feingold or Barrett in recall” it said in another on May 25, 2011.
But Walker’s prospects have improved since then.
“Walker faces fight of his life in Wisconsin recall election” a PPP headline read on Feb. 28 of this year.
And with a recall election is slate for June 5, the über-liberal website the Daily Kos cited a PPP survey showing Walker with a 5-point lead on one likely Democrat in the recall election and a 7-point lead on another.
Merely reporting numbers or trying to create a narrative?
It’s nearly impossible to determine the motivations of any political polling organization — especially one that has made no qualms about avowing its support for one party over another (Jensen said he was “absolutley rooting” for the Democrat in a 2010 North Carolina Senate race and PPP’s founder Dean Debnam reportedly donated $2,400 in the 2010 Democratic primary).
But in a world full of political reporters seeking to devine which hometown candidate is winning and which one is losing — combined with an avid segment of the electorate on the left and the right with no shortage of passion — organizations like PPP and Rasmussen have found an eager audience for their automated polling and instant data.
But maybe the audience needs a dose of caution and skepticism.
In October of 2010, a New York Times political blogger reported about “a Democratic official, who asked not to be identified, who told me that he had largely discounted using automated polls after discovering systematic differences between them and polls conducted by live interviewers.”
Caveat emptor, New Mexico voters, caveat emptor.
Update 4/27: In regards to the Scott Walker numbers and PPP, on April 16 the Washington Post wrote about the support for Walker in a new PPP survey now that the recall election is nearing.
The key lines:
According to Public Policy Polling, Walker leads Milwaukee Mayor Tom Barrett (D) 50 percent to 45 percent and former Dane County Executive Kathleen Falk (D) 50 percent to 43 percent.
That’s a shift in PPP polling from February, when Barrett had a slim lead over Walker while Falk tied the incumbent.
Pollster Tom Jensen explained that it was the organization’s first to use a likely voter screen for this race.
That’s what Republican critics refer to when they allege that PPP is trying drive a narrative in the early stages of an election cycle.
Does that prove that PPP is guilty of that? No, but as the Washington Post story says, it’s worth mentioning.