May 29, 2008
A single photograph provides a static view of its subject. But when we connect a series of images together and shuffle through them, as with a flip-book or the frames of a movie, what we see corresponds more closely to the real world.
The same is true of opinion polls. One poll gives us a snapshot of the public at a particular moment in time. Having two, three, or more polls on the same topic reveals the range of public opinion, its consistencies and its volatilities.
Thus, Wednesday’s new California Field Poll on marriage equality attitudes — the second poll to be publicly released on this topic since the state Supreme Court’s May 15 marriage decision — expands our knowledge about Californians’ reactions to the ruling and their views about overturning it in November.
For supporters of marriage equality, the news is good. The Field Poll results may mean that last week’s LA Times poll — which itself provided hopeful signs for the November election — actually painted too pessimistic a picture.
Field Poll Findings
The new poll included two questions about general attitudes toward marriage equality. First, repeating a question that has been included in the Field Poll since 1977, respondents were asked:
“Do you approve or disapprove of California allowing homosexuals to marry members of their own sex and have regular marriage laws apply to them?”
- 51% Approved
- 42% Disapproved
- 7% had no opinion
This marks the first time ever that a majority of Field Poll respondents has supported marriage equality. In a 2006 survey, by comparison, 44% approved while 50% disapproved.
The poll also included a question — similar to the one asked in last week’s LA Times poll — to which respondents indicated which of three different statements about legal recognition of same-sex relationships most closely resembles their own view:
- 45% selected “gay and lesbian couples should be allowed to legally marry”
- 32% selected “gay and lesbian couples should be allowed to form civil unions or domestic partnerships, but not legally marry”
- 19% selected “there should be no legal recognition of a gay or lesbian couple’s relationship”
- 4% had no opinion.
Regarding the pending state ballot initiative, the poll randomly divided the sample and asked each group of respondents a slightly different version of the question.
VERSION A: “Do you favor or oppose changing the California State Constitution to define marriage as between a man and a woman, thus barring marriage between gay and lesbian couples?”
- 40% Favor
- 54% Oppose
- 6% DK
VERSION B: “There may be a vote on this issue in the November election. Would you favor or oppose having the State Constitution prohibit same-sex marriage, by defining marriage as only between a man and a woman?”
- 43% Favor
- 51% Oppose
- 6% DK
The Margin of Error
When considering the results of this (or any) poll, it’s important to keep in mind that the percentages reported above are estimates of how many people in the entire population hold these opinions. Although the specific percentage is the best guess that can be derived from the data, the poll actually provides a range of percentages within which the correct value for the population probably lies. This range is described by the poll’s margin of error.
For most of the Field Poll marriage questions, the margin of error is about 3 percentage points. Thus, what the poll really tells us is that there’s a high likelihood that somewhere between 48% and 54% of California voters currently approve of marriage equality, while somewhere between 39% and 45% disapprove.
For the two versions of the question about amending the California constitution, the margin of error is slightly larger (because each question was asked of only a portion of the sample). Between 50% and 58% of voters oppose the ballot measure as it was described in Version A, while 36% - 44% favor it. And between 47% and 55% oppose the B Version, while 38% - 48% favor Version B.
Note that the ranges of those favoring and opposing Version B overlap. This is what’s meant by a virtual tie or a statistical dead heat.
The results for both versions indicate that the anti-marriage initiative may be in trouble before it even qualifies for the ballot. To quote from a previous post:
The LA Times article noted that “ballot measures on controversial topics often lose support during the course of a campaign” and, for this reason, “strategists typically want to start out well above the 50% support level.” According to Susan Pinkus, the Times Poll Director, “Although the amendment to reinstate the ban on same-sex marriage is winning by a small majority, this may not bode well for the measure.”
To get a better sense of the California electorate’s current views, it’s useful to compare the Field Poll results with last week’s Los Angeles Times survey on the same topic.
Both polls found that attitudes toward marriage equality and toward the ballot initiative differed substantially according to age, party affiliation, political ideology, and geographic region. Younger respondents, Democrats, liberals, and residents of Los Angeles and the San Francisco Bay Area were more likely to support marriage equality, while older respondents, Republicans, conservatives, and residents of the Central Valley and other rural areas tended to oppose it. In both polls, the attitudes of moderates and independents were closer to those of Democrats and liberals than to those of Republicans and conservatives.
The polls differed, however, in their findings about gender. The Field Poll didn’t report men’s and women’s attitudes toward the ballot measure separately, suggesting that they didn’t differ. But in the Times poll, likely women voters overwhelmingly said they would support the initiative — 62% to 29%. (Men were evenly split — 44% to 44%.) As I explained in another post, this pattern puzzles me. Past research on heterosexuals’ attitudes toward sexual minorities and policies affecting them has revealed a fairly consistent gender gap, with heterosexual women less prejudiced and more supportive of gay civil rights than heterosexual men. And previous polling in California has found that marriage equality receives more support from women than men.
In addition to the gender difference, the polls differed in many of their key numbers, even when we take their respective margins of error into account. In the LA Times poll, between 32% and 38% favored marriage equality (compared to 42% - 48% in the Field poll) and 26% - 32% opposed all legal recognition of same-sex couples (compared to 16% - 22% in the Field Poll). And the Times estimate of voters who would support the November ballot initiative is between 50% and 58% — nowhere near the ranges for the Field Poll’s Version A (36% - 44% favored it) or Version B (38% - 48% favored it).
Such differences can be caused by a variety of factors. For example, data collection for the Field poll spanned a considerably longer period - 10 days, compared to 2 days for the Times poll. This may mean that its sample was ultimately more representative because having more time means more opportunities to reach sample members who initially weren’t at home.
The wording of poll questions can also affect response patterns, especially among those who don’t hold strong opinions or haven’t thought about the issue extensively. In this regard, perhaps some differences resulted from the fact that the Field Poll asked about general attitudes toward amending the California constitution, while the LA Times poll asked about respondents’ actual intention to vote for or against the measure.
What Does It Mean?
As new data become available, we’ll get a better sense of whether the LA Times or the Field Poll better describes the California electorate. It will be especially interesting to see new results from polling by the Public Policy Institute of California, which has tracked Californians’ attitudes on this issue throughout the decade and found increasing support for marriage equality. In their 2007 survey, the PPIC found that likely voters were evenly divided, with 46% supporting marriage equality and 48% opposing it.
For now, the results from the two available polls suggest the November ballot initiative is currently in trouble — capturing only a slight majority (LA Times), losing outright (Field Poll Version A), or being too close to call (Field Poll Version B).
This doesn’t mean supporters of marriage equality should be complacent. Elections are decided by actual ballots, not survey responses, and backers of the initiative will be working hard to get out the vote among those most likely to endorse the amendment in November.
Indeed, as an antidote to overconfidence, it’s instructive to consider the Field Poll’s batting average in predicting the outcomes of California ballot initiative contests. In 2000, Field correctly predicted that Proposition 22 (the antigay Knight Initiative) would pass. However, it seriously underestimated the extent of support for that initiative. The last Field Poll before election day found that 53% of voters supported Prop 22, but it ultimately passed with 61% of the vote.
Needless to say, a similar undercount could be occurring in this year’s polling.