How Much Does Abortion Polling Actually Matter?

A bortion is an infamously difficult issue to poll, made all the more challenging by the fact that the issue was confined to public-opinion polling — rather than explored through public political debate — since the 1973 Roe v. Wade decision. On other hot-button political topics, pollsters have been able to regularly test the validity of their survey findings against the political reality of how people vote. When votes don’t match up with polls, they can go back to the drawing board and rethink their methodology. (Or at least they’re supposed to.) Because Roe effectively removed abortion from the democratic process, there have been very few comparable opportunities over the past half century for analysts to discuss the political implications of the issue in anything other than abstract terms. Dobbs changed that.

On the evidence of the polls, Americans occupy almost every imaginable position on abortion. On the one hand, a Gallup poll conducted in May found that 55 percent of Americans identify as “pro-choice,” versus just 39 percent who identify as “pro-life.” The next month, a Pew poll found that 61 percent of U.S. adults believe that “abortion should be legal in all or most cases,” compared with just 37 percent who believe that “abortion should be illegal in all or most cases” — and the month after that, Pew found that 57 percent of Americans disapproved of the Supreme Court’s overturning of Roe v. Wade.

On the other hand, a September 2021 poll found that Americans are evenly split on whether abortion should be legal in all or most cases (a combined 49 percent said legal in all or most cases; a combined 49 percent said legal only in some cases or not legal). And the popular support for Roe is belied by the fact that a significant majority of Americans don’t actually understand what overturning it meant: Public-opinion surveys, including those showing sweeping support for Roe, have routinely found that Americans incorrectly believed that overturning the decision would ban abortion nationwide rather than returning the issue to the democratic process. Incoherently, a 2015 Vox poll found that “of people who say abortion should be legal only in cases of rape, abuse, or if the woman’s health is at risk, 53 percent oppose overturning Roe v. Wade.”

Things get even more complicated when we break things down by trimester. As Alexandra DeSanctis wrote in May, “a Gallup poll from a few years back found that only 28 percent of Americans favor allowing abortion in the second trimester, and only 13 percent of Americans favor allowing it in the last three months of pregnancy — compared to 60 percent who would allow it in the first trimester.” That’s a discrepancy that has been consistently shown across a number of surveys in recent years: A June 2021 AP poll found that “61% of Americans say abortion should be legal in most or all circumstances in the first trimester of a pregnancy,” but “65% said abortion should usually be illegal in the second trimester, and 80% said that about the third trimester.” Again, there’s an enormous amount of ideological incoherence to all this. As FiveThirtyEight noted earlier this year, Roe “legalizes abortion up to the last few weeks of the second trimester, underscoring another important factor here: It’s not just that Americans hold contradictory opinions on abortion. Many Americans simply don’t know a lot about abortion — including its legality and accessibility where they live.”

America is now witnessing the first real-world test in 50 years of how abortion legislation plays in the political arena. The picture is markedly different from what little we could glean from decades of polling. This is particularly relevant regarding the 15-week federal ban on abortion that Senator Lindsey Graham (R., S.C.) introduced earlier this month. Polling consistently shows that such a measure should be popular — a Trafalgar poll on Graham’s bill conducted last week found that 59.1 percent of Americans support the 15-week ban, whereas 40.9 percent oppose it. But how much can this kind of polling actually tell us about how Americans will vote?

It’s far too early to draw sweeping conclusions about what kinds of abortion restrictions are political winners — and which are losers — just a couple of months after Roe was overturned. The first national test of how voters feel about the issue will come this November, which is why the 2022 midterms “might end up being The Most Important Election for pro-lifers,” as Dan McLaughlin suggested last week. But we’re already seeing signs that polling and other forms of hypothetical political analysis can tell us even less about what Americans think about abortion than we might have previously thought. For decades, pro-life groups — for entirely legitimate and understandable reasons — have cited polls showing overwhelming support for restrictions on second- and third-trimester abortions to argue that Americans are not aligned with the maximalist pro-abortion-rights position. That may well be true, but political reality is also much more complicated than the granular positions that Americans express when asked by pollsters.

First of all, it’s important to note that only a relatively small minority of Americans rank abortion as a top issue. A Monmouth poll released in early July found that just 5 percent of Americans said abortion was their top issue, with 33 percent saying it’s inflation and 15 percent naming gas prices. Similarly, an August Gallup poll found that 8 percent of Americans rank “abortion issues” as their top priority. (Both Monmouth and Gallup did find that the number of Americans who ranked abortion as their top issue had increased from previous surveys.) To be fair, an NPR poll from earlier this month found that 22 percent of Americans ranked abortion as their top issue, and previous surveys that expanded the question to include voters who view abortion as relevant to their voting decision rather than their top issue showed larger returns — a 2020 Gallup poll found that “47% say abortion issue is one of many important factors to their vote,” with 24 percent saying a candidate must share their abortion views, versus 25 percent who said it was “not a major issue.” But even those surveys don’t crack a majority, and the most ideologically committed abortion voters remain less than a quarter of the sample polled.

In other words, while committed abortion voters tend to hold more ideologically coherent views on the issue, most Americans probably don’t think about the topic all that much. It’s reasonable to conclude, then, that Americans whose views are reflected in abortion polling for the most part aren’t very attached to the specific stance they take in answer to pollsters; those views are fluid, and could very well be subject to change if abortion were to become a more salient topic in a particular election. Whether or not these voters change in the pro-life or pro-choice direction likely depends on the context in which the issue arises.

The gap between opinion polling on abortion and political reality was visible in the failure of the recent Kansas referendum on an amendment to remove the right to abortion from the state constitution. A 2018 Public Religion Research Institute poll of abortion opinion by state found that views on the issue in Kansas were similar to those in other red states that have passed sweeping abortion restrictions in recent years: Just under half of Kansans (48 percent) believed abortion should be legal in most or all cases, comparable to the Republican-controlled states that border Kansas — 47 percent in Oklahoma, 49 percent in Missouri, and 43 percent in Nebraska. (Colorado, Kansas’s Democratic-controlled neighbor to the West, sat at 56 percent.) In 2020, 56 percent of Kansans voted for Donald Trump, who had by then appointed the three Supreme Court justices who would help to overturn Roe. And polling on the constitutional amendment itself showed the pro-life side with a narrow lead going into the August 2 vote: Just weeks before, a poll found that 47 percent of Kansans supported the measure and 43 percent opposed it, with the remaining 10 percent saying they were undecided. But the referendum ended in a devastating loss for pro-lifers, with Kansans voting down the amendment by a margin of 59 to 41 percent. For now, abortion will remain legal through the 22nd week of pregnancy in the Sunflower State.

There are any number of explanations for why the Kansas referendum failed. But the main takeaway, as Alexandra DeSanctis noted, is that “supporters of abortion effectively won the messaging battle.” The failure of the referendum is not a sign that the pro-life cause is doomed in America. But it is a cautionary tale about the explanatory limits of political data when it comes to the issue.

Where will voters for whom abortion is not a top or well-defined issue ultimately draw the line on abortion? The average American’s view of the issue is likely reactive — it depends on how it’s presented to them in the moment. Voters will probably balk in response to later-term procedures; people know, instinctively, that it’s wrong to dismember a child. But at least initially, they’re also likely to view more restrictive bans as a step too far. That seems to be part of the reasoning behind Senator Graham’s bill. As I argued back in May, “given that voters are not entirely aligned with either side, the party with a better message will be the one that ultimately prevails.”

To win over voters open to persuasion, Republicans have work to do in conveying the gruesome nature of abortion. What that means is that Republicans need to be willing to tap into the electorate’s most visceral instincts about the issue. Pro-lifers need to remind Americans of what it is abortion does. Particularly in the post-Roe era, the confused picture that emerges from abortion polling is even less useful than it was in the past. If there’s one thing that the polling can tell us, it’s that most Americans really don’t know what to think about the issue. The task, for the pro-life movement, is to change that.