Man, I thought we were done with 2010 election surveys. Since we didn’t have any exit polls this year, KSTP and SurveyUSA (who must have some extra cash lying around) did some post-election temperature-taking.
I’ll get to KSTP’s headline in a sec, but if you’ve ever wondered whether people lie to pollsters, there’s some pretty strong circumstantial evidence here.
The 1,400-person poll’s very first question: “Did you vote in Tuesday’s election for Minnesota’s governor?” 81 percent said yes.
The actual figure: 58 percent.
Usually, when pollsters poll, they’re dealing with an unknown outcome. Here, we have an actual result. So how could SurveyUSA be so far off?
Could be sampling error, but that’s only 2.1 points plus or minus. (That means the “I voted” result should’ve been between 79 and 83 percent 95 times out of 100). Other design effects — the absence of “cell phone only” voters, high refusal rates — could have played a part.
But given SUSA’s general accuracy this cycle, a simpler scenario is that people fibbed. Researchers call it a “socially desirable response” — you’re likelier to tell a stranger (or, in this robo-poll, a stranger’s recorded voice) that you did your civic duty.
SUSA surveyed 1,400 Minnesotans. In the poll, 1,136 said they voted. But 58 percent of 1,400 equals 812. If the difference was all lying-related, perhaps 300 of the 1,400 bore false witness.
Interestingly, this poll was nearly on the money in mirroring the governor’s race result. It showed Dayton 45, Emmer 44, Horner 9, when the actual stands at 44/43/12.
But that gets to my other beef here: KSTP’s headline that “Horner drew more votes from Dayton.”
The survey did show Horner voters would’ve picked Dayton over Emmer 37-30 had Horner not been on the ballot. As reporter Tom Hauser noted, the question had a “huge” margin of error — 9.9 points plus or minus. That’s because the sample size is so small, just 103 respondents who voted (or lied about voting) for Horner.
I understand that when you take a poll, you want to use the results. But sometimes, the sample is so pathetic that it’s better to bury certain answers than headline them.
Of course, the election is over and no voter will be influenced by this information. Still, I’d hate for political junkies or historians to conclude “Horner helped Emmer” based on this very thin reed.