I’m not the only one who has noticed the apparent desperation from the Jane Norton campaign. Fellow RMA and PPC blogger Don Johnson, frequently criticized for showing favoritism to the former lieutenant governor, now says that a Norton victory would make her “the miracle candidate of the year.” Among other things, Johnson describes a newly released Norton campaign poll, conducted by Public Opinion Strategies (POS), showing their candidate with a 39-33 lead as “unbelievable.”
On Tuesday afternoon, upon invitation from the Norton campaign, I joined their press teleconference to discuss the results of this poll — which contradicts recent polls from Magellan and SurveyUSA showing Buck with double-digit leads. (While Nate Silver’s numbers rank SurveyUSA as the third most accurate political polling firm out of 63, POS rates 60th. The in-state firm Magellan was not included.)
As articulated by Norton campaign spokesman Josh Penry, the POS poll assumes greater participation in the August primary because of high Republican voter intensity and a mail-in ballot election. Simply put, Norton’s numbers indicate the broader the electorate, the greater share of votes she gets.
I certainly don’t doubt this analysis. The question will be whose turnout model proves to be correct. While I wouldn’t go as far as calling the POS poll “unbelievable,” from where I stand, odds are currently against enough voters turning in their ballots to put Jane Norton over the top.
Penry told me he believes the other two polls likely are flawed for depending on lists of Republican voters who participated in the 2006 and 2008 primary elections. (He said their polling suggests that especially younger people, and females, who didn’t vote in those elections are much more inclined to support Norton.)
So I contacted both SurveyUSA and Magellan to inquire about their voter lists and screens used in their respective polls that show Ken Buck in the lead. Magellan uses a current list of Republican voters and asks participants how likely they are to vote in the August primary. This self-selecting method yielded their sample of 1,026 likely voters. I left a message with SurveyUSA, but don’t really expect to hear back.
Of the three polls, I tend to give Magellan the most credence because it’s the only one to do all of the following:
- Has a reasonable method for identifying likely voters [It's debatable whether this also applies to either POS or SurveyUSA]
- Has the largest sample and the lowest margin of error [POS has the smallest sample, in no small part because it doesn't rely on automated survey techniques]
- Shows a significant share of undecided voters (26%), which logically comports with the state of the campaign — given the timing and level of media attention thus far [SurveyUSA does not]
- Has a geographically and demographically balanced sample of voters (e.g., 51 Female, 49 Male) [SurveyUSA does not, though as I pointed out, it makes only a minor difference]
- Overlaps in margin of error with another of the two polls [POS does not]
Is it a stretch at this point to say Ken Buck has majority support and a 16-point lead in the primary? Very likely.
Is it a stretch at this point to say more likely primary voters are prepared to back Jane Norton than her opponent? Almost certainly.
All of my usual thousand words to end up where I began a couple weeks ago after the Magellan survey came out:
Given the significant amount of voters “probably” supporting their respective candidates and a quarter of those polled expressing indecision, making more than two-thirds of the electorate at least somewhat open to be swayed, it’s fair to say the race is still very much up for grabs.
But if Ken Buck can continue his momentum by reporting a significant increase in fundraising through the end of this month, all signs will show him as the come-from-behind prohibitive favorite for the last month of the primary campaign.