A poll giving a slender lead to opponents of the proposed amendment to the Minnesota Constitution banning same-sex marriage was released Thursday by a team of faculty members and students from St. Cloud State University.
Fifty-one percent of likely voters plan to vote against the proposal to insert a same-sex marriage ban into the constitution, while 44 percent support it.
A rarity in this election cycle, it was conducted independently and using traditional, time- and labor-intensive methods.
The poll also found eroding support for the other question on this year’s ballot, the proposed voting amendment. Thirty-nine percent would reject the measure and 55 percent of likely voters plan to vote for it.
Undecided voters and those who said they would skip voting on the ballot questions — effectively a “no” vote in Minnesota — would not tip the balance, according to the survey of 600 Minnesota adults conducted Oct. 15-21. The margin of error is 5 points.
The SCSU effort is the first since the barrage of 11th-hour TV and radio ads, which often move polling needles, began airing.
The numbers jibe with those produced by other recent polls, which have found the voting amendment still far ahead but with eroding support and the gay-marriage ban trailing slightly.
But unlike many of the voter surveys sparking headlines recently, this one used gold-standard methodologies and appears to be quirk-free. Indeed, its principal investigators, political science professors Steve Wagner and Stephen Frank, have released a 25-page document [PDF] describing — in excruciating detail — how the numbers were obtained.
The important bits: The survey used a sample that was balanced on multiple levels. Forty percent of the numbers were for cell phones, the same percentage as Minnesota households without landlines. Live interviewers — trained students working under close supervision — alternately asked to speak with men and women and the oldest and youngest adult in the households that were called.
Voting amendment
| All respondents | Likely voters | |
|---|---|---|
| Vote for | 52% | 55% |
| Vote against | 36% | 39% |
| Not vote | 8% | 3% |
| Don’t know | 4% | 3% |
Respondents were asked not just their political identification and how likely they are to vote, two key elements in a poll, but a series of questions designed to ferret out actual likelihood and leaning. Also asked were questions about age, education, household income, county of residence, gender and religion.
Many surveys circulating in recent months have had local and national public opinion geeks scratching their heads because the crosstabs, the data subsets pollsters use to gauge a poll’s credibility, haven’t added up. Some have shown young voters, who typically support same-sex marriage in large numbers, voting in favor of the ban, for example, while others have confounded expectations regarding gender and rural vs. urban residence.
Marriage amendment
| All respondents | Likely voters | |
|---|---|---|
| Vote for | 41% | 44% |
| Vote against | 49% | 51% |
| Not vote | 7% | 3% |
| Don’t know | 3% | 2% |
MinnPost’s David Brauer and I have both written critical stories about the discordant surveys, so before pronouncing this one a to be solid and independent, I contacted longtime local polling guru Rob Daves, who teaches survey research and the University of Minnesota’s Humphrey Institute and is the principal of Daves & Associates.
As a neutral pollster himself, Daves declined to venture an opinion about the numbers themselves but cheerfully parsed the pages of fine print concerning methodology.
“I’m glad to see a respected university poll using both cell phones and landlines in their research, compared with some of the potentially less accurate methodologies we’ve seen in the Minnesota media in the past week or two,” he said. “Support could change dramatically in the next two weeks because of the barrage of media ads and news that will center around the two amendments.”
So does this mean we can call the ballot questions and turn our collective attention to post-election topics? Hardly.
Two more items of interest regarding the SCSU poll: Regarding the marriage amendment, the numbers are almost exactly what they were when the scholars asked voters the same question one year ago yesterday.
“The overall results showed that 44 percent of respondents believed that the Minnesota constitution should be amended and 47 percent believed that it should not be amended,” they reported, while “9 percent of respondents either refused to answer or said that they were not sure.”
Other polls have shown the proposed marriage ban starting out with a majority but losing support steadily in recent months.
Lastly, this year’s survey team also asked voters whether they were confused by the wording of the proposed amendments. Critics have asserted that the language describing the voting amendment in particular is unclear.
Regarding each of the amendments, a third of respondents told the SCSU team they found the questions confusing. What does that say about the validity of their answers about how they’re going to vote? Stay tuned.
The long and sordid history of polling in Minnesota
Why anyone pays attention to St. Cloud State poll, or Rob Daves who used to run the Minnesota Poll, or the Minnesota Poll is a mystery to me. St. Cloud missed the 2010 Dayton-Emmer race more than anyone. The Minnesota Poll has a long and sordid history of overstating DFL support as this graph by the Washington Post demonstrates:
http://www.washingtonpost.com/wp-srv/politics/documents/MinnesotaPoll_kennedycampaign071706.pdf
Here are the Minnesota Poll results and the actual totals for the past four gubernatorial elections, the most important election in the state, and the one it should be the easiest to get right:
Final Results
2010 Dayton 43.6, Emmer 43.2, Horner, 11.9.
2006 Pawlenty 46.7, Hatch 45.7, Hutchinson 6.5.
2002: Pawlenty 44, Moe 36, Penny, 16.
1998: Ventura 37, Coleman 34, Humphrey 28.
Final MN Poll
2010 Dayton 41, Emmer 34, Horner 13.
2006 Pawlenty 40, Hatch 45, Undecided 7
2002 Pawlenty 35, Moe 32, Penny 16.
1998 Coleman 30, Humphrey 35, Ventura 27.
Compare the other polls
The SCSU poll shows a slightly narrower margin in favor of Obama than other recent public polls in Minnesota, not a wider one. That doesn’t sound to me like a pro-Democratic skew.
And your argument falls apart when you call out the SCSU poll by presenting numbers from past STrib Minnesota polls. It gets even worse when the one individual you imply with a common past of Minnesota Poll and SCSU turns out not to be among SCSU’s current researchers. Go check out the SCSU survey site. Daves’ only connection with the SCSU poll is to have offered his punditry for this particular MinnPost article.
Crosstabs?
Does anyone know whether the crosstabs are available for this survey? Or results on any of the other questions?