The Problem With Polling Abortion

Polls serve an essential function in American democracy, but they have inherent limits we would be wise to acknowledge

This is difficult for me to admit as someone who makes a living conducting and analyzing polls, but we might need to think hard about what polls are really contributing to the public debate on abortion.  

Polls play a crucial role in our debates over public policy. They reveal to journalists, elected officials, and policymakers the public’s preferred policy outcome, as well as how important certain issues are to them. They are an indispensable resource for those seeking to represent the collected public will.

But they can also be used to oversimplify complex views on complex subjects. On no other public policy question are the limitations of polling more evident than on abortion. One of the most dramatic examples of this shortcoming can be found in Gallup’s question about abortion identity. The question requires respondents to identify either as “Pro-life” or “Pro-choice.” It suggests that Americans are evenly and intractably divided over abortion.

But we know that this does not actually reflect the true contours of public opinion. More than other issues, abortion lends itself to multiple competing perspectives and conditional opinions. Americans judge behavior differently depending on the circumstances. And even a seemingly simple and straightforward question such as Gallup’s becomes much more complicated when respondents do not have to choose between mutually exclusive options. One survey allowed respondents to either select or reject both pro-choice and pro-life labels. Many did. Only half of Americans exclusively identify as pro-life or pro-choice. The rest identify with both of them or neither.

That Americans’ attitudes on abortion are complicated is not exactly revelatory. In a 2018 survey, Americans were asked whether they think abortion is a complex or simple issue. Most Americans believe abortion is complicated. And the people who believe the abortion issue is “simple and straightforward” are most likely to express the most extreme views. 

But surprisingly, when you dig a little deeper, our differences do not seem quite so stark. A recent report, “How Americans Understand Abortion,” based on in-depth interviews with over 200 Americans, finds that Americans eschew easy categorization in favor of thoughtful engagement on the issue. The report's authors note that Americans are earnest and thoughtful in their approach to thinking about abortion. “Americans—by and large—do not approach abortion with callousness, but with sensitivity and a recognition that it is a tough issue. Attitudinal differences about abortion’s morality and legality do not diminish the weightiness of abortion’s impact in real life, on real people.” Most Americans wish there were fewer abortions, and even those who support the legality of abortion acknowledge it’s not a desirable outcome.” This is perhaps the ultimate irony of the abortion debate—we’re not as divided as it appears.

At the very least, in describing public attitudes on abortion, we should resist the urge to characterize the entire landscape of opinion using a single question. Many Americans who support the legality of abortion still express moral reservations about it. And many pro-life Americans generally believe there are certain exigencies where abortion should be allowed. We have more overlapping identities, beliefs, and concerns than are evident from any one question.

One of the downsides of polling on complex issues like abortion is that it is necessarily reductive. Ambivalence is difficult to convey in forced-choice questions and is often portrayed as contradiction or inconsistency. Polling not only reduces complicated feelings to multiple choice answers, it encourages us to segment people into discrete and often opposing groups. But most Americans do not approach difficult issues in this way. On the issue of abortion especially, we need to offer more space for Americans to express their fears, share their experiences, and convey deeply held beliefs than is possible in a poll.