Polls on Abortion Can Be Misleading

Print More

Professor Anna Greenberg

(WOMENSENEWS)–The biggest news about people’s views on the legality of abortion is that almost nothing has changed for the last three decades: Americans in large numbers support the legality of a woman’s right to choose abortion, just as they have since Roe v. Wade was decided in 1973.

Interviews with experts and reviews of polling data reinforce this conclusion, although it is also true that information is muddled in many news reports because many journalists don’t understand the intricacies of polling. In fact, questions that include such labels as pro-choice or pro-life often fail to elicit accurate, meaningful answers.

Although many polls on abortion are conducted by candidates and organizations, the Gallup Organization in Princeton, N.J., is the main public pollster to consistently and continuously survey this question over time. One of the primary methodologies of polling is to use the same question repeatedly so that answers in different time periods can be compared. The experts quoted here used the Gallup polls as an example to examine the shortcomings and hazards that are part and parcel of polling in general and polling on the abortion issue in particular.

The Gallup Organization, a private, commercial polling corporation with $300 million in annual sales, released a new analysis of abortion data on Jan. 22, a nearly yearly ritual in order to snare headlines on the anniversary of the Supreme Court’s Roe v. Wade decision. Except for a single question, the analysis was based on prior polls by Gallup.

Gallup’s results showed that only 19 percent of Americans believed that abortion should be illegal in all circumstances, based on a poll in March and April 2000. This data compares to 22 percent who responded the same way in April 1975, in its first poll.

Gallup’s January report failed to place front and center its only new research on abortion, which showed that nearly two-thirds of Americans–63 percent–are either content to leave abortion laws as they are or want abortion laws to be less restrictive. This question was asked Jan. 10 to 14, 2001, but the results were to be found only on page four of the six-page report.

Over Two-Thirds Oppose Constitutional Amendment Outlawing Abortion

The polled public’s rejection of anti-abortion positions is further indicated by results from October polling by Gallup, which showed that more than two-thirds of Americans–67 percent–opposed “a constitutional amendment to overturn the Roe vs. Wade decision and make abortion illegal in all states.” With only a glancing mention of this result at the outset, the report morphed the “two-thirds” into a mere “majority.”

“The basic bottom line is that people want the status quo,” said Carroll Doherty, report editor of the Pew Center for Research, who reviewed the data.

Even abortion scholar Karlyn Bowman of the notably conservative American Enterprise Institute, where other scholars include Newt Gingrich and Robert Bork, said new data has not altered results of a guest scholar paper she wrote for Gallup in January 2000. “What is striking is that a quarter century of debate has not significantly changed attitudes on abortion,” she wrote.

Despite this information, one point in the Gallup report seemed to indicate a shift on abortion attitudes and caused considerable confusion because of a delicate and error-prone blending of labels with viewpoints.

“With respect to the abortion issue, would you consider yourself to be pro-choice or pro-life?” This is a question to which Public Agenda online, a website sponsored by Cyrus Vance’s nonprofit education organization in New York, attaches a red flag warning, saying findings may be “misleading, unstable or easily misinterpreted.”

‘Pro-choice,’ ‘Pro-life’ Label Questions Elicit Misleading Responses

Gallup reported these findings high in its January report. Frank Newport, editor-in-chief of The Gallup Poll, noted in an e-mail response to a related e-mail poll inquiry, “The abortion controversy is one which has focused heavily on labels, on both sides, and we have always attempted to be neutral and unbiased.” But journalists who relied on the article written by Gallup’s Joseph Carroll–with top placement on the label question–were led astray.

The first problem for reporters was in thinking that the Gallup results indicated a recognizable shift in opinion. The findings–47 percent selected the “pro-choice” label and 45 percent said “pro-life”–were collected in October 2000. But only three months earlier, in July, 50 percent of the people called themselves “pro-choice” and 40 percent “pro-life.” No January polling was reported.

Gallup announced that people were “evenly split” and that the percentage agreeing with the pro-choice label had diminished since 1996, an arbitrary date selected to make this point. It ignored that the July figures were also the largest-leaning pro-choice gap since 1997.

“We’re talking about changes of 56-48-53-47-51.These are not big changes, considering how difficult it is to measure these kinds of attitudes in general,” said Anna Greenberg, a professor on polling techniques at the John F. Kennedy School of Government at Harvard University. “These are not massive swings of opinion. So I’d say it’s relatively stable,” said Greenberg.

But an even more significant problem appears to be that the labels are meaningless and bewildering to the general public. In focus groups, people who select either a “pro-choice” or “pro-life” label may take very contrary positions when specifics are probed, said Molly O’Rourke, vice-president of Peter Hart Associates, a polling firm in Washington, D.C.

Non-Expert Media Wrongly Reported Even Split on Abortion, Based on Poll

O’Rourke said news reports of the Gallup results worried her until she realized the question was about labels and not viewpoints–“apples and oranges,” she said. The label information can be useful to groups preparing messages, but do not reflect people’s opinions, she said.

“I find ‘pro-choice’ and ‘pro-life’ are irrelevant, the most irrelevant things that exist,” said Jeff Pollack, president of the Global Strategy Group Inc., a New York firm that conducted an in-depth national survey for the Othmer Institute of Planned Parenthood of New York in June 2000.

Gallup did not emphasize this distinction in its report, even using “label” and “sentiment” interchangeably at one point. Some media followed suit. Reuters online news reported that Americans were divided over abortion and cited the Gallup results on the label question.

The normally reliable Kaiser Daily Reproductive Health Report, an online summary of abortion news, based a headline on the label question and wrote that it showed “Americans are ‘evenly split’ between those who support and those who oppose abortion rights.”

A Salon.com story went even further, tartly chastising abortion rights supporters for casting President George W. Bush outside the mainstream on abortion, using as evidence the label question of Gallup. “Foes of abortion have quietly gained ground in recent years,” the story said, ignoring the rest of the polling analysis.

“Many journalists,” said Greenberg, “don’t know anything about polling data or how to analyze it.”

Big Picture Is That Opinions Don’t Fluctuate Much From Year to Year

When public pollster Louis Harris and Associates last conducted a poll of 1,000 adults in January 1998, the 25th anniversary of Roe v. Wade, it also found almost the same responses as it had in 1973. Only 17 percent of the people in the 1998 Harris poll said that there were no circumstances under which they would “favor permitting a woman who wants one to have an abortion.” This compared with 20 percent of the people who answered the same way in 1985.

Twenty-three percent of those surveyed said they would favor abortions “under all circumstances,” and 48 percent said under “some circumstances.” This compares with 1985 data in which 26 percent selected “all” circumstances and 53 percent said “some circumstances.”

One thing that the polls show about abortion in a big-picture view, said Lois Timms-Ferrara, associate director of the Roper Center for Public Research at the University of Connecticut, is that people have definite and fixed opinions. “They don’t fluctuate a lot from year to year,” she explained.

The biggest pools of people actually float somewhere in the middle on abortion, explained Robert J. Blendon, a professor of public health and political analysis at Harvard University School of Public Health and Kennedy School of Government. The majority believe, he said, that abortion should be legal, but are willing to accept some restrictions.

Another finding is that there are two distinct groups in the middle: those who are willing to accept some restrictions, even though they support the legality of abortion, and others who will agree to exceptions, even though they are not strong supporters of legalized abortion.

In the March Gallup poll, 51 percent chose the middle category of “legal under only certain circumstances,” with 28 percent agreeing with the more liberal statement that abortion “should be legal under any circumstances.” These numbers have also changed little over the years: In 1975, 54 percent placed themselves in the middle and 21 percent in the more liberal group. Divining the nuances of how the middle might shift and change, depending on the circumstances or language, is what keeps many pollsters in business.

In the end, the public polling is driven by an elite political debate, says Greenberg from Harvard, and tells little about the invisible issues and real experiences with abortion–whether providers are available, services accessible, fees affordable or what people know about those concerns.

“These things are just as important to the whole Washington debate as Roe v. Wade,” she said. “We almost never see public pollsters ask.”

Cynthia L. Cooper is a free-lance journalist in New York, and writes frequently about reproductive rights issues.

Comments are closed.