As predicted, the SDGOP spin machine has reversed its embrace of the latest Survey USA poll on South Dakota's big political races.

On Tuesday, when Pat Powers thought the Survey USA Senate race results affirmed his snarky theses about Weiland being an awful candidate and heading for third place, he gleefully bruited the poll in multiple posts. He called the results fact and chortlingly concluded they meant the death of the Democratic Party.

But then Pat realized what the poll results really say:

  1. His man Mike Rounds is the awful candidate, the big-money erstwhile foregone conclusion who now can't buy his way past 40%.
  2. The Independent Larry Pressler, whom Powers mercilessly and classlessly mocks at every turn, has (in John Tsitrian's priceless words) managed to "stumble into factor-hood" and has a campaign on fire.
  3. Corinna Robinson and Angelia Schultz can poll within striking distance of their supposedly popular Republican counterparts without lifting a finger.
  4. And the biggest vote getter on the ballot will be Initiated Measure 18, the Democratic measure to raise South Dakota's minimum wage.

"Pretty soon Pat Powers is going to figure out that he shouldn't be crowing about the latest Survey USA poll," I wrote yesterday. "Expect some fake analysis from Pat any moment now explaining why we can't trust Survey USA."

Cue fake analysis:

I’ve reached the conclusion based on these wildly varying results that I’m not buying it. And I suspect Democrats are over represented in the poll.

Part of the problem is, according to its methodology, it’s a push button poll combined with Internet polling [Pat Powers, "Anyone Buying the Survey USA Polling Anymore?" Dakota War College, 2014.09.12].

Get Pat started on math, and things are bound to go wrong.

Sorry, Pat. Democrats were not over-represented. Look at the grey line at the bottom of each question's breakdown, showing "Composition of Likely Voters." Then call your pal and former patron Secretary of State Jason Gant to compare those numbers with South Dakota's voter registration totals as of September 2, the day before Survey USA started calling:

Survey USA Voter Reg
GOP 50% 47%
Dem 33% 34%
Indy 16% 19%

Survey USA over-represented Republicans, not Democrats. Survey USA under-represented Independents. A more representative sample would have put Rounds even lower and Weiland and Pressler even closer to pulling off the upset of the decade.

Pat expresses concern that Survey USA used online polling. Pat links to the wrong methodology page, which doesn't mention the Internet. The poll page itself explains the methodology better:

Cell-phone and home-phone respondents included in this research: SurveyUSA interviewed 775 South Dakota adults 09/03/14 through 09/07/14. Of the adults, 674 were registered to vote. Of the registered voters, 510 were determined by SurveyUSA to be likely to vote in the 11/04/14 general election. This research was conducted using blended sample, mixed mode. Respondents reachable on a home (landline) telephone (88% of likely voters) were interviewed on their home telephone in the recorded voice of a professional announcer. Respondents not reachable on a home telephone (12% of likely voters) were shown a questionnaire on their tablet, smartphone, or other electronic device [Survey USA, "In Solid Red South Dakota, Republican Rounds is Held to 39%...," 2014.09.12].

12% of respondents answered online. That 12%, accessed from lists purchased from various vendors, showed the same support for Weiland, two points less for Rounds, and three points less for Pressler than the landline respondents. Pat does not provide any analysis about why those results are less trustworthy. But I can tell you that 12% under-represents the number of South Dakotans who have no landline.

And if Pat is going to reject polls that combine online polling and phone-button responses, he's going to have to throw out all of his old posts trumpeting Rasmussen Reports. Read their methodology: they do pretty much the same thing, with robocalls augmented by a sample of landline abandoners with an "online survey tool."

I wonder if Pat's even trying any more. He doesn't offer analysis, just wishful thinking frosting his sponsors' press releases. Why bother, Pat?