Unlikely Voter

Poll Analysis and Election Projection

Checking in on the polls after the last caucuses

When I left for CPAC, Mitt Romney had just won the Nevada caucuses 50-21 over Newt Gingrich, numbers reasonably in line with the last poll, by Public Policy Polling.

In DC I found out Rick Santorum came out of nowhere and did well in Minnesota, Colorado, and Missouri. Let’s see if the polls caught it.

I have been and will continue to be critical of PPP when the firm asks stupid, trollish questions in Republican primary-related polling. But their late polling, when the election is near, tends to drop the partisan foolishness and start projecting reasonable results. So it’s not surprising to me at all, contrary to what I see in some circles, that their results have managed to be reasonable. Well, reasonable for caucus polling, anyway, which is tough (or even impossible?) to do.

Anyway, PPP had Missouri at Santorum 45, Romney 32, Paul 19. The actual results were Santorum 55, Romney 25, Paul 12. Some of the difference could be explained by a late Santorum rush, but they did get the ordering right.

PPP was also the only firm to do Colorado. In this state, they weren’t so accurate. Of course, in Missouri Santorum was already ahead, but in Colorado there was a sudden Santorum gain. Romney 37, Sanotrum 27 in the final poll became Romney 35, Santorum 40. Note: Romney’s support was predicted. Santorum gained support from Newt Gingrich, who went from 21 in the poll to 13 in the actual results. This lends support to the theory that Rick Santorum is becoming the “anti-Romney.”

On the other hand, there’s Minnesota. The final PPP poll was frankly nothing like the results. The actual results were Santorum 45, Paul 27, Romney 17, Gingrich 10. The final poll was Santorum 33, Romney 24, Gingrich 22, Paul 10. Was this a case of organization paying off? Or just a bad poll?

So, only one pollster hit these three states with their surprising wins, and we got three different evaluations. One state PPP got right, one PPP got almost right, and one PPP got badly wrong. Sadly after all that we can only conclude two obvious facts: Polling caucuses is hard bordering on worthless, and Rick Santorum had a good week.

Comments

One Response to “Checking in on the polls after the last caucuses”

  1. “I have been and will continue to be critical of PPP when the firm asks stupid, trollish questions in Republican primary-related polling. But their late polling, when the election is near, tends to drop the partisan foolishness and start projecting reasonable results”

    Well, can you please then tell everyone to stop repeating stupid trollish PPP polss that are not “late”? For example, the obviously wrong PPP national poll from a couple days ago that said that Santorum was up 15 points over Romney (we know it is obviously wrong because there have been FOUR polls released since PPP, all using the same dates for polling, that showed the race within 3 points).

    Republicans – you included – are giving waaaaay too much credence to a Democrat polling firm. They ought to be completely shunned by Republicans.

Write a Comment

Comments are closed.