Behind the Shifting Poll Numbers on Abortion

By Jodie Allen, Thomas Jefferson Street blog

Experienced pollsters often find the trend in responses over time as interesting—or even more interesting—than the absolute numbers recorded. This is especially the case when dealing with controversial or novel topics, the sort of question that may tend to make respondents more likely to give what they see as the socially or morally acceptable answer rather than what they truly think.

So it caught the attention of the experts at the Pew Research Center when they noticed in an April 2009 poll that the proportion of American adults saying they thought that abortion should be legal in all or most cases had declined by 8 percentage points from the level recorded as recently as August 2008. Whereas in the 2008 poll those favoring legal abortion outnumbered those in opposition by 54 percent to 41 percent, now the two sides were essentially tied at 46 percent to 44 percent.

Wondering if the latest finding was simply a one-time anomaly—even in the best designed and administered polls these do crop up from time to time—or a real change in public attitudes, they decided to follow up the question in two August national surveys to see both if the results were stable and also whether demographic patterns in the responses suggested a possible explanation for sudden shift.

The August surveys were conducted in two waves, (Aug. 11-17 and 20-27) yielding a nationwide sample of 4,013 adults, 18 years of age or older. Adding in the April sample produced a rich lode of 5,500 respondents, ample enough to slice and dice in multiple ways. For purposes of comparison, surveys from 2007 and 2008 were combined to yield a total sample of more than 14,000. Moreover, as the Pew Research Center's Director of Survey Research Scott Keeter notes, in designing the August surveys they were extra careful not to place the abortion questions behind other questions that might influence responses, so as to "get the cleanest reading possible."

The overall findings from the total sample closely matched the April results with the proportion choosing "legal in most or all cases" at 47 percent, a 7-percentage-point drop from a year earlier, while the number saying abortion should be illegal in all or most cases rose by four points to 45 percent—producing a virtual tie between the two sides.

The report of the findings, released on October 1, drew widespread attention in not just the press, but also some criticism from other pollsters. For example, Gary Langer, director of polling at ABC News, expressed skepticism noting that other polls, including ABC's and a summer Washington Post poll, did not show a trend away from support for legal abortion. "Our June 21 result almost exactly matched our long-term average in polls since 1995, 56-42 percent," Langer wrote, adding: "It is notable that Pew's gotten mid-40s twice this year. Nonetheless, I'd sure like more consistency across polls to call it a trend—especially given other recent data." For example, he pointed to another ABC June finding that fully 60 percent said they'd wanted then-nominee Sonia Sotomayor to vote to uphold Roe v. Wade, were she to be confirmed as a Supreme Court justice.

Still, Langer acknowledged, his polls this year had showed some shifting—both rightward and leftward—on social issues so, "We may be in a period of some changeability as people reassess these positions with the Democrats in power, one of the suggestions Pew's making."

And, indeed, the Pew Research report concluded that both the timing of the shift—coming after the election of Democrat Barack Obama as president—and its distribution among demographic and political groups "suggests that the election of a pro-choice Democrat as president may be a contributing factor."

The findings are bolstered, Keeter observes, not only by the very large size of the samples from which the trend is derived but also by the fact that the observed "pattern of change made some sense—that it didn't seem to be random noise." For example, it is telling that no fall off in support for legal abortion was seen among Obama's most supportive groups: African Americans, young people and those unaffiliated with a religion. At the same time, declines were most pronounced among Republican-leaning independents (a group that now included many former Republicans) and church-going Catholics but also among Democratic men—though not Democratic women, says Keeter.

Other factors also bolstered Pew Research's conclusion that the trend was real, though the report also stressed that few people say abortion is a critical issue today (15 percent) compared with 2006, when 28 percent described abortion as a critical issue facing the country.

Read the report and judge for yourself—and stay tuned for further readings on the trend.

• Check out our political cartoons.
• Become a political insider: Subscribe to U.S. News Weekly, our digital magazine.
• Follow the Thomas Jefferson Street blog on Twitter.