When the Doug Burgum campaign released an internal poll they claimed showed a “statistical tie” with opponent Wayne Stenehjem I expressed some skepticism. Based on reports I got from SAB readers who were questioned for this poll it wasn’t so much a preference poll as a messaging poll.
That is a poll where the respondents are asked carefully crafted questions representing talking points the campaign might use. The respondents are usually asked about their candidate preference more than one time during the questioning so that the impact of the messages being tested can be measured.
The Burgum campaign told me that the numbers they were touting – 44 percent for Stenehjem and 40 percent for Burgum – were from a candidate preference question asked “prior to any probing questions about Doug or Wayne.”
If true, that would put some weight behind the number, but should we just take the campaign’s word for it? Also, they say their was weighted (which is to say manipulated so that it better represents the electorate being measured) pre-survey, but where are the details on that?
That’s typically released with a preference poll, but it wasn’t released here.
Anyway, the Fargo Forum editorial board pressed Burgum about the poll in an editorial meeting, asking him to release all the qeustions, and according to a report from Helmut Schmidt he refused. Though he said he’d think about it.
The polling, done April 5-6 by Wilson Perkins Allen Opinion Research, had Burgum as a favorite in a hypothetical ballot among 40 percent of those polled. Stenehjem had 44 percent, Paul Sorum 1 percent and undecideds were at 15 percent. The margin of error for the poll was plus or minus 4.4 percent.
“We’ve released what we’re going to release, but it’s wishful thinking by the opposition to think that that poll was unbalanced or just positioning,” Burgum said. “Because this is the poll which we’re making our decisions off of, and it was done by (a) highly qualified, expert firm in political polling and I believe the results because … I saw the polls that they’ve done earlier. I’ve seen the movement. I know the methodology. It was clean.”
“Some people are saying ‘Oh, it was a push poll.’ That’s inside baseball,” he said.
I suspect that if the roles were reversed, and the Stenehjem campaign was touting results from a messaging poll favorable for them, Burgum might be singing a different tune.
I don’t think it’s “inside baseball” at all to point out that the purpose of messaging polls isn’t to measure candidate preference but to measure the effectiveness of messaging.
This doesn’t mean that the poll results Burgum is touting are inaccurate. They could be spot-on. It’s just hard to reach that conclusion based on the provenance of a poll that a) isn’t decide to measure candidate preference and b) isn’t released in a transparent way to the public anyway.
And I doubt the Burgum campaign will release the full poll. That would mean revealing what messaging questions they asked, which might tell us what sort of messages the Burgum campaign is weak. And it would tip their hand for the sort of messaging which may be effective against Stenehjem.