Margin Of Error: Breaking down the polls

Margin Of Error: Breaking down the polls

What NC's polls tell us on gay marriage

Posted October 14, 2014

Logo for the Margin of Error blog.

— Do North Carolinians oppose gay marriage?

This question seems timely given recent events. Last week, the US Supreme Court allowed lower court decisions to stand that made it illegal to ban gay marriage in certain states. As a result, same-sex marriage is now legal in North Carolina, despite a state laws and a constitutional amendment to the contrary.

Let's try to answer our question by starting with one claim opponents of gay marriage have made. Amendment One, which prohibited gay marriage, passed with 61 percent support in 2012. Thus, it appears a majority in NC opposes gay marriage. For example, Shelly Carver, who is Senator Pro Tempore Phil Berger’s spokeswoman, said Republican leaders are going to intervene "to defend the will of more than 60 percent of North Carolina voters who cast ballots defining marriage as between one man and one woman in our state constitution."

I don’t agree that vote means a majority of North Carolinians opposes same-sex marriage. I agree that in 2012, a majority probably did oppose same-sex marriage. Public Policy Polling (PPP), for example, found 55 percent would vote for Amendment One in their final poll before the vote. Yet, the vote on Amendment One took place over two years ago, and the political landscape for gay marriage has rapidly changed since then. And even in 2012, other polls such as Elon’s suggested support for same-sex unions combined with same-sex marriage was the majority opinion in 2012, and that support for Amendment One was partly based on confusion over its meaning.

Second, we care about what all North Carolinians think, not just voters. The vote on Amendment One took place in the primary elections, when voter turnout is at its lowest for statewide and federal offices. Rather than random draw of citizens, primaries attract citizens who are not representative of the broader population. In the 2012 primaries a little more than 2 million citizens voted on the amendment, which was 34 percent of eligible voters.

Third, surveys in NC and around the country find growing acceptance of same-sex marriages, suggesting that with more time majority support is all but inevitable. Not only are people changing their minds more so in the direction of support, but also younger people are increasingly part of the adult population and also the most likely demographic to support same-sex marriage. Demographic trends ensure support will continue to rise.

Growing support for gay marriage

The two most recent surveys that North Carolinians’ opinions about gay marriage are evenly split, although their results depend partially on who is being asked. Since elections attract more polls, and most election polls only care about what registered voters or "likely voters" think, representative polls on the topic are scarce.

The Elon Poll of registered voters from September finds that supporters are slightly more prevalent than opponents, 45 percent versus 43 percent, although the opinion gap is statistically indistinguishable at high levels of confidence. Likewise, the American Insights Poll conducted at the same time found registered voters were evenly divided, 46 percent vs 46 percent. Among likely voters, opponents were more frequent, 48% vs. 44%. I have not seen a poll on the same topic with a representative sample of adults in North Carolina, though I would guess it would measure higher support for gay marriage compared to these results by about 2-3%.

It was interesting to see that both the Elon Poll and American Insights reached nearly identical conclusions despite asking the question quite differently. Elon simply asked, “Do you support of oppose gay marriage?” Conversely, AI asked, “Do you think marriages between same-sex couples should or should not be recognized by the law as valid, with the same rights as traditional marriages?” Despite these differences, their findings were largely parallel, suggesting people are reporting firmly held opinions impervious to changes in question wording.

I was also intrigued by another question AI asked about how best to decide the debate over gay marriage. AI asked, "Regardless of your own preference on the same-sex marriage issue, do you believe North Carolina law defining marriage should be determined by voters or courts?" Just 26 percent said the courts should decide the issue, while 62 percent said voters. Despite the instructions to disregard their own preference, I looked and found that just 51 percent of supporters of gay rights thought voters should decide the outcome, but 77 percent of opponents wanted voters to decide. It appears each side has accurately surveyed the political landscape and knows which option is most likely to secure their preference. Although, in this case, I'd argue surveys show that opponents are racing against time.


Please with your account to comment on this story. You also will need a Facebook account to comment.

Oldest First
View all
  • Doug Hanthorn Oct 17, 2014
    user avatar

    While this article does mention the fact that Amendment 1 was voted on in a primary election and that typically primaries are low turnout affairs, it neglects to mention that the Democratic primary had no important races and thus many Democrats decided or neglected to vote. The Republican primary did have very important races and thus Republicans turned out in much higher numbers. You can argue that if Democrats had considered Amendment 1 important, they would have turned out just to vote against it, but nonetheless, the Republicans who put the amendment on the primary did so deliberately to give it the best possible chance of passing as opposed to putting it on the general election ballot where it would have received the greatest voter turnout to decide the issue. It was a disingenuous ploy on the part of Republicans through and through and at least deserves a mention here.

  • skeeter II Oct 15, 2014

    How the asked question is worded influences the response received. Remember when the child asked "you don't want me to stay overnight with my friend, do you ?" The question suggest the answer wanted.

    I believe that it would be meaningful if the questions asked and the responses received should be released along with the poll results. We could then, individually, decide if the "results" are acceptable to us.

    Also whether the entire "universe" is available to be "sampled" is a question. Many young people do not have land-lines, only cell phones or tablets. The makeup of those contacted will also influence the results received. If the majority of those contacted were by land-line, the results would be different than if the majority were contacted through their cell phone or tablet.

    A large margin of error could nullify the results -- declaring that the results should not be relied upon.