Moments ago, KCTS9, in cooperation with the Center for Survey Research at the University of Washington, released the results of this year’s Washington Poll. The Washington Poll is a widely respected electoral survey directed by Dr. Matt Barreto, who serves as an associate professor at the UW in Seattle. It is conducted annually, and because its findings have historically correlated more strongly with actual election results, it is considered to be a very credible poll.
We at NPI are not overly fond of polling. Obsession with poll numbers has, in our view, distorted the traditional media’s coverahe pf politics and elections. Horse-race style coverage (in which the focus is on who’s winning and who’s losing) has become more predominant, displacing stories produced as a result of actual newsgathering (which admittedly costs more to do, but is consequently much more valuable).
Nevertheless, well-directed and thoughtfully conducted public opinion research deserves our attention, and the Washington Poll is certainly such a survey. Unlike other polls, it only comes out once a year, and it is conducted by live callers calling both mobile lines and landlines (as opposed to automated dialing software).
The 2012 Washington Poll was in the field beginning October 1st and ending two days ago (Tuesday, October 16th). Interviews generally were about a half hour in length. The samples are composed of seven hundred and eighty-two registered voters and six hundred and forty-four likely voters were interviewed. Margins of error for these samples are 3.5% and 3.9%, respectively.
Here’s a look at the results.
For president, the poll finds Barack Obama with a very comfortable lead. 51.8% of registered voters are certain or planning to vote for Obama, while only 41.3% are certain or planning to vote for Romney. The numbers are not so different for the sample of likely voters… 51.9% Obama, 42.9% Romney. Less than four percent of each group of respondents characterized themselves as undecided.
Obama’s lead in Puget Sound tops 60% according to the crosstabs, and he has an even stronger lead statewide among women voters (54.6% to Romney’s 37.9%) and young voters (57.3% to 34.1%).
The governor’s race is much closer. Jay Inslee has a narrow lead, but his lead is within the margin of error. Among registered voters, Inslee leads 47.9% to 44.7%. Among likely voters, Inslee leads 48.3% to 45.1%.
The breakdown shows that Inslee is ahead in Puget Sound, while McKenna is ahead everywhere else. McKenna has a slight edge among male voters but it is at a slight disadvantage with female voters. Republicans are slightly more likely to vote for McKenna (95.5%) than Democrats are to vote for Inslee (89.5%).
Inslee’s strongest supporters are young people. 55.6% of those between the ages of eighteen and twenty-nine are for Inslee, with only 36.4% for McKenna. McKenna, on the other hand, does not have a comfortable lead with the other age groups. He leads Inslee with each age group above thirty, but not by very much.
The U.S. Senate race looks like it will be a blowout. Maria Cantwell is way ahead, with 58.3% certain or leaning towards voting for her. Baumgartner’s share of the vote is only 34.8%. Among likely voters, Cantwell’s share drops to 57.7% and Baumgartner’s increases to 35.4%, but it’s not much of a difference. Cantwell is ahead in every area of the state except the east, where she and Baumgartner are roughly even. Young people are overwhelmingly in Cantwell’s corner (65.7% to Baumgartner’s 26.1%), but she leads with every age group.
The Washington Poll also looked at ballot measures. Respondents were asked about each of the three initiatives on the ballot, along with Referendum 74 (marriage equality).
The poll found that support for Tim Eyman’s I‑1185 is only at 53.6%. Opposition stands at 31.2%. Considering how few resources have been put into the NO campaign, that’s a fairly encouraging result. Only 37.8% described themselves as certain they were voting for I‑1185. An additional 13.3% described themselves as yes, but could change. A further 2.5% are undecided, but lean yes.
23.8% of respondents, meanwhile, are firmly opposed. 5.7% are opposed but could change their minds, and 1.7% lean no. 14.6% describe themselves as undecided. For an issue that’s been on the ballot before, most recently two years ago, that seems like a pretty high number of undecided voters. There is clearly an opportunity for progressives to bring the yes vote for I‑1185 down in the final weeks of the campaign.
The numbers for likely voters are not much different, so I won’t go over them.
I‑1240, the latest attempt to bring charter schools to Washington, appears to be in trouble. The total yes vote is below the fifty percent mark, which indicates that voters have doubts about the measure. Only 34.1% are certain yes. 30% are certain they’re voting no. 10.6% are planning to vote yes but could change their minds, and 7.3% are planning to vote no but could change. A further 2.8% and 1.9% are undecided but lean yes and no, respectively, while 12.6% are simply undecided.
As with I‑1185, the numbers for likely voters are not much different, so I won’t go over them.
I‑502, the initiative to legalize marijuana, has a respectable lead among registered voters (50.9% total yes, 40.8% total no). I‑502’s overall lead among likely voters is not as impressive (47.1% total yes to 40.1% total no), but the percentage of likely voters who are firm nos (31.8%) is several percentage points lower than the sample of registered voters (35%), while the number of respondents who described themselves as certain to vote for I‑502 was about the same.
The percentage of voters undecided also differed between the samples. 7.6% of registered voters say they’re undecided, but that rises to 11.7% for likely voters.
Finally, the Washington Poll looked at Referendum 74 (marriage equality). For Referendum 74, the results were adjusted to account for what Barreto and his team call “social desirability bias”. Here’s what they mean by that:
Ballot initiatives on marriage equality tend to over-report the margin of victory for those campaigns. The Washington Poll’s estimate of yes votes on Referendum 71 was within the margin of error, but we under estimated the percent of no votes.
To rectify this, we included two items on this year’s poll designed to measure social desirability bias.
The first question asked individuals if they lied on the survey; the second asked if any topics made them uncomfortable. Both groups of people in the Yes on 74 and the No on 74 camp reported lying at rates that did not differ from the sample average. Those who reported that they would not vote on the issue and those who were undecided reported lying at rates higher than the sample average.
This relationship remained statistically significant for those who were undecided even after running a logit model to control for religiosity, age, ideology and partisanship. Additionally, those who reported that they weren’t voting on the issue reported being more uncomfortable with the questions about lesbians and gay men.
A higher percentage of people in this group were more uncomfortable with these questions than those who voted reported voting no.
2.15% of those who said that they were voting for Referendum 74 said that they were uncomfortable with the questions about gay men and lesbians.
In light of these results, we adjusted the point estimate. The result is a conservative estimate of social desirability bias. We kept the Yes on 74 and No on 74 columns stable except for the yeses who reported discomfort on the sexual orientation question. We then moved the undecideds and those who reported that they would not vote on the issue into the No on 74 column because they reported lying more than those who were decided. This resulted in a point estimate of 52.9 to 46.6, a 6.3 point difference.
The adjusted results for Referendum 54 show the “Approve” side ahead, 52.9% to 46.6%.
Voters who live west of the Cascades are far more supportive of marriage equality than voters who live to the east, which is not surprising. A plurality of voters in every age group are supportive of marriage equality, but young voters — and voters approaching retirement — are far more enthusiastic than seniors and middle-age voters. Take a look:
|Age 18–29 (born late 1980s or 1990s)||68.2%||26.1%||5.7%|
|Age 30–44 (born 1970s, 1980s)||50.2%||39.1%||10.7%|
|Age 44–65 (born 1950s, 1960s)||59.1%||34.3%||6.6%|
|Age 66+ (born before the fifties)||48.6%||41.5%||9.9%|
These results are remarkable. Look at which age groups are closest to one another. Voters born in the late 1980s or 1990s (Generation Y, the millennial generation) are very enthusiastic about marriage equality; that’s to be expected. But the age group they’re closest to on this fundamental question of civil rights is not the thirty to forty-four year olds. It’s the forty-four to sixty-five year olds — Washingtonians born in the fifties and sixties — who are nearest to them. A surprising 59.1% of voters in that age group are in the Approve column.
The most senior age group — those over sixty-six — divide somewhat along the same lines as the Generation X group, though there are slightly more people in the Reject column and slightly fewer in the Approve column.
These results suggest that Referendum 74 is on track to be approved in a few weeks. The divisive Reject campaign is kicking into high gear, but they have a tough hill to climb. Support for marriage equality has been steadily building over the years, driven by young people, who are clearly shifting public opinion in a healthy, progressive direction.