A survey of 1,191 likely voters by Public Policy Polling has the amendment ahead, 58-38 percent, while a poll of 1,001 by SurveyUSA has the amendment up, 58-36 percent.
Each survey included in the question the language that will appear on the ballot: "Marriage between one man and one woman is the only domestic legal union that shall be valid or recognized in this State."
North Carolinians will vote May 8.
The two polls differ dramatically from an Elon University survey that has gotten considerable attention in the state but used different wording in its surveys. Elon surveyed 534 adults and found North Carolinians opposing the amendment, 62-31 percent. But the survey's question stated the issue in the negative: "Would you an amendment to the North Carolina Constitution that would prevent any same-sex marriages, domestic partnerships, or civil unions?"
Nationwide, polls have shown that people are more likely to support a marriage amendment if it is stated in the positive -- as the ballot language does and as the Public Policy Polling and SurveyUSA polls did -- than if it's stated in the negative, as the Elon poll does. Questions that include the word "prevent," as the Elon survey does, also tend to decrease the level of support for an amendment. The word "any" also may have led to a lower level of support for the amendment, since it underscores the negative tone of the question.
In a Tweet earlier this year, Public Policy Polling's Tom Jensen called it a "fatal flaw" for Elon University not to use the "exact ballot language" that voters will see when they enter the booth May 8. Public Policy Polling's survey of Maine in 2009 was the only one to correctly predict residents there would reverse a law that had legalized gay "marriage."
The Public Policy Polling survey was conducted March 23-25; the Survey USA poll, March 16-20; and the Elon University poll, March 26-29.
Compiled by Michael Foust, associate editor of Baptist Press.
Copyright (c) 2012 Southern Baptist Convention, Baptist Press www.BPNews.net