The story of the anti-Romney poll calls into Iowa and New
Hampshire that I wrote
about yesterday gets stranger and stranger. Here is the lead of the story reported
last night by the Chicago Tribune's
The GOP presidential campaigns of
Mitt Romney and John McCain-rocked in different ways by a highly negative
"push poll" targeting Romney's Mormon faith-demanded Friday that the
New Hampshire attorney general investigate who is behind the tactic. The
attorney general's office said it was investigating the phone calls.
Again with feeling: This particular set of calls sounds more
like an ethically questionable "message testing" survey than a classic "push
poll." See my post
from yesterday for more details on that issue or the clarification
released last night by the American Association for Public Opinion Research
(AAPOR - full disclosure: I serve on AAPOR's executive council).
An interesting twist to the story, according to Zuckman's
story, is that a "New Hampshire law requires all political ads-including
phone calls-to identify the candidate behind the effort, or at least the
candidate who is being supported."
I went looking for more details about the questions asked on
the calls, and the most detailed report comes from State Representative Ralph
Watts, a Republican from Adel, Iowa. He taped a radio
interview with Radio Iowa that you can listen to online.
Here is the way he describes the interview (my transcription):
It started out like a lot of
telephone polls do these days. They wanted to know if I was a caucus goer, and
whether I was a regular voter and all that usual stuff. And then it progressed
into questions about Mitt Romney, and specifically about the Mormon Church.
The first one, I guess, was
innocent enough. It asked a question whether I would be more or less likely to
vote for Mitt Romney because he's Mormon. Well, I guess that's a fair question,
but not necessarily a pertinent question. And then it went on to talk about the
philosophy of the Mormon Church. Would I be more or less likely to vote for
Mitt Romney based on some of the tenants of the Mormon Church?
This telephone interview went on
for about 20 minutes. The last half of it were questions directed, they were in
a more positive light and they were directed toward John McCain. They asked a
question, what if I knew that McCain had some 330-some carrier landings and was
a Navy pilot would that make me more or less likely to vote for him. If I knew
that John McCain were a prisoner of war in Vietnam would it make me more or
less likely to vote for him. Then there was a whole series of questions about
John McCain that were very favorable questions about John McCain. It would have
led one to believe that John McCain were behind the poll, but that would have been too obvious.
And I've done some checking myself
and [with] some people, and I'm convinced that John McCain had nothing to do
with it. Who actually did it, there you don't know.
What Watts describes starts
out with typical political survey questions, then shifts to a long series of
negative arguments about Mitt Romney followed by a long series of positive
arguments about John McCain. The length of the interview and type of questions
is indicative of a "message testing" survey. Ordinarily, that pattern would
suggest a survey conducted by someone supportive of McCain looking for the best
ways to promote their candidate and to most effectively tear down Romney. However,
between the red-hot spotlight of presidential politics and the incendiary
nature of questions about Romney's religion, there is nothing ordinary about
It is tempting to try to use the facts reported by Watts and
other respondents to logically deduce the identity of the sponsor of the calls.
But readers need to remember two things about contemporary "push poll" stories:
First, respondent memories are often imperfect. They will
often exaggerate some details and omit others. Consider that in a 20-minute
interview, a pollster can typically ask 60 to 80 questions. In the description
above, however, Representative Watts specifically recalls just a half dozen or
so questions. My point here is not to challenge his story, only to suggest that
the reports we have are so far cover only the most memorable details. We may be
missing some useful context.
Second, and probably most important, keep in mind that
accusations of "push polling" have become a fact of life for campaign
pollsters. Since virtually all campaigns in both parties now conduct "message
testing" surveys, and since most reporters reflexively (and erroneously)
describe any report of a negative question on a survey as evidence of "push
polling," pollsters have grown accustomed to being so accused. Unlike the calls
involving the Democrats I wrote
about earlier in the week, these calls have all the hallmarks of a
professional survey, including the length of the questionnaire and the use of a
well-regarded call center. So given the intense media spotlight on Iowa and New
Hampshire and the explosive nature of questions about Romney's Mormonism, my
guess is that the pollster that designed this survey assumed the calls would
lead to a "push poll" story. Perhaps that assumption is a part of their
And that's what makes it impossible to try to deduce from
the available facts the campaign or interest that was behind the calls. As
Representative Watts says, the positive questions about McCain are almost "too
obvious" as a ploy intended to implicate McCain is the sponsor.
So who is behind these calls? I haven't a clue, but the
story gets stranger and stranger.
Follow Mark Blumenthal on Twitter: www.twitter.com/MysteryPollster