Noem over Jackley by 11? Poll Provenance Suggests Otherwise

Last month on live radio, in front of thousands of South Dakotans, I said Marty Jackley will beat Kristi Noem in the Republican gubernatorial primary 58–42.

On Friday, Noem released a poll saying I’ve got the spread wrong by 27 points. The scant history of the guys behind the Noem poll suggests they have the spread wrong by 27 points.

According to the Noem poll press release sent from Washington, D.C., Kyle Robertson is running something called Rockbridge Strategy. Robertson  worked for Mike Pence in 2010. Robertson speaks of the poll numbers, gathered by Republican campaign firm Advantage Inc., as a sign that Noem has lead that “may force Jackley” to buy lots of negative ads, “a risky strategy as one would assume his image would take a hit.”

Stop right there: I smell some narrative manufacturing. We all know that Jackley and Noem didn’t each raise seven figures to just run nice videos of themselves in checked shirts riding horses all spring. Once they get their nice-guy/gal promos out of their system, they’ll start whacking each other on TV and radio. Jackley and Noem know the negative attacks are coming; Robertson’s assertion sounds like an effort to pre-emptively deflect whatever negative attack Jackley opens with—lies about dad’s estate tax? abandoning House Ag?—by making readers think Noem’s awesomeness forced him to be nasty.

Mississippi campaigner Justin Brasell is running Noem’s campaign—poorly, one might conclude, given his botched attacks on Jackley over travel budgets and trial lawyers.

Brasell and Robertson tag-teamed in opposite roles in the Kentucky governor’s race in 2015. Robertson was managing the campaign of Hal Heiner, one of four Republicans on the primary ballot. Brasell was consulting via Triumph Campaigns. On March 26, 2015, Brasell’s Triumph Campaigns found 601 Kentucky Republican voters to give Robertson’s candidate, Hal Heiner, a fifteen-point edge over his closest competitor:

1. Hal Heiner: 33%
2. James Comer: 19%
3. Matt Bevin: 12%

On May 19, 2015, the Kentucky Republicans who voted in the primary upended those results:

1. Bevin: 32.91%
2. Comer: 32.89%
3. Heiner: 27.10%

Over two months, a 21-point spread cooked up by Brasell favoring Robertson’s candidate turned out to really be a six-point loss. 27 points off.

Maybe Brasell and Robertson get better results when they switch roles, as they have for Noem, but if their 2015 accuracy matches their current accuracy, we would expect to see over two months an 11-point advantage turn to a 16-point deficit… which is the spread I predicted last month.

Jackley’s manager Jason Glodt notes the past collaboration of Brasell and Robertson and dismisses the Noem poll as an “April Fools joke.”

We should always view with suspicion polls announced from out of the blue by campaign consulting firms. They are usually more interested in shaping reality than in reflecting it. Nonetheless, there are some fun numbers in the data… which we’ll parse in the next post!