How Do Americans Really Feel About Abortion Rights?
[Note: The following newsletter was written on May 12, 2022, 12 days before the U.S. Supreme Court overturned Roe vs. Wade. In this newsletter I describe some of the reasons why it’s difficult to characterize Americans’ attitudes about abortion. Although these difficulties remain, I want to emphasize two of the conclusions I drew: “The majority of Americans believe that abortion should be legal, at least under some circumstances….” and “The majority of Americans want the Supreme Court to uphold Roe v Wade.” I would hate for these simple conclusions to get lost among my somewhat fussy discussions of methodological uncertainties and other nuances.]
I doubt that anyone can describe how Americans "really" feel about a complex, politically sensitive, and intensely personal topic such as abortion rights. Still, it's worth trying to understand public opinion, given the Supreme Court's imminent reversal of Roe v Wade, and given the Senate's rejection yesterday of a bill that would've guaranteed women the legal right to abortion while blocking numerous restrictions on access.
In this newsletter, I'll address two questions: 1. Why is it difficult to characterize public opinion on abortion rights? 2. What can we say with confidence about public opinion on abortion rights?
1. Why is it difficult to characterize public opinion on abortion rights?
In other words, why do we see inconsistencies, disagreements, and uncertainties in descriptions of America public opinion on this topic? In this section I’ll describe two culprits: survey methodology and media coverage.
(a) Survey methodology.
Given that public opinion is a moving target, our best source of information consists of surveys with rapid turnarounds conducted by news organizations (e.g., the New York Times), governmental agencies (e.g., the CDC), and large, nonpartisan groups (e.g. Gallup, Pew, NCOR). However, owing in part to the need for speed, all of these surveys are methodologically limited.
Consider for a moment one of the best and most influential examples, the General Social Survey (GSS). This nonpartisan survey includes six questions that begin with the phrase: “Please tell me whether or not you think it should be possible for a pregnant woman to obtain a legal abortion if...?" The six clauses that complete these questions are: "if pregnancy endangers the woman’s health", "if the woman has become pregnant as a result of rape", "if there is a strong chance of a serious defect in the baby", "if she does not want any more children", "if her family is very low income and cannot afford more children", and "if she is unmarried and does not want to marry the man. Recently, the GSS added a seventh question: "....under any circumstances."
Later I'll discuss the GSS findings. Here I just want to emphasize two key limitations:
(i) Respondents contradict themselves. Specifically, NCOR has found that 20% of respondents who indicated, in response to the seventh question, that abortion should be legal under any circumstances, also stated in response to at least one of the previous questions that abortion should not be legal in that particular circumstance. It would be acceptable if a small percentage of people exhibited this sort of inconsistency – survey respondents do sometimes become distracted, inattentive, or confused. However, 20% is a lot of people. This suggests a problem with the survey questions (it's unclear at present what the problem might be).
(ii) Much is missing from this survey. For example, the GSS treats abortion as a black-and-white issue – i.e., either you want it to be legal or you don’t. However, for many women, the pressing issue is not legality per se, but rather legal restrictions on access. A woman may have the legal right to an abortion, but, depending on the state, she may face waiting periods, counseling requirements, and bans on insurance coverage, and she may have to travel hundreds of miles owing to restrictions on medical providers. If you're poor, but your insurance doesn't cover abortions, and you couldn't take enough time off work anyway, then the fact that abortion is legal is a moot point. It's legal, but, for legal reasons, it's inaccessible. The law says you can do it, but, in effect, the law prevents you from doing it.
The GSS abortion questions are in the process of being revised. Meanwhile, no survey or combination of surveys seems methodologically adequate. For example, a Gallup poll, administered at least once per year to a nationally representative sample since 1988, does address many of the access-related omissions I mentioned, along with others (e.g., the question of how late during pregnancy abortion should be a legal option). However, Gallup doesn't ask all of these questions each year. For many questions, the most recent data are from 2018. For some issues (waiting periods, parental consent, partial-birth procedures) the most recent data are from 2011.
One other limitation of existing surveys is what can be called insufficient granularity. None of them capture important but subtle distinctions that are known to exist in peoples' attitudes (a common problem in survey research.) For example, Gallup's 2018 survey asks whether abortion should be legal or not under a variety of circumstances pertaining to the mother (e.g., pregnancy that endangers her life, or that results from rape or incest) as well as the baby (e.g., mental disability, or life-threatening illness). These questions are asked with respect to first trimester abortions, and, separately, with respect to third trimester ones. Remarkably granular. And yet, other survey data, as well as qualitative interviews, tell us that some people's views would be shaped by considerations such as the projected degree of risk to the mother, the type of mental disability the baby has, and whether the baby's illness would be treatable. And, people care about issues not mentioned in the survey, such as the financial burden to the mother of raising a baby with a mental disability or illness. In short, the survey can't capture the entire scope of peoples' thinking.
(b) Media coverage.
Some news sources are ideologically biased to the point of administering biased surveys or intentionally misrepresenting published data. But I believe that even organizations that attempt to be fair-minded sometimes misrepresent public opinion, owing to at least two "myths".
Myth 1: Attitudes toward abortion rights are completely politicized.
In other words, everything that an individual or organization says about abortion rights will reflect their political ideology. Democrats are pro-choice, Republicans are pro-life, and Independents are one or the other depending on which way they lean.
Political ideology does inform attitudes toward abortion rights, but the connection is less clear than often assumed. Here's a nice way of summing it up that I came across this week: "For most of us, abortion is many things and not all of them fit neatly into boxes".
Where did I read this? In a Fox News article, written by a woman who wants to preserve Roe v Wade and ensure that abortion is "safe, accessible, and legal" for all women. So, yes, if you read Fox News, you will often see pro-life perspectives, but not exclusively.
More broadly, political affiliation is not synonymous with attitudes toward abortion rights. A typical finding, reported last year by NORC, is that 40% of Democrats oppose completely unrestricted legal abortion (i.e., access to an abortion for any reason), while 29% of Republicans support completely unrestricted legal abortion.
In short, political affiliation only "predicts" attitudes toward abortion rights in the narrow sense that Democrats are more likely than Republicans to support unrestricted abortion, while Republicans are more likely than Democrats to oppose it. Journalists who equate political party with views on abortion rights will be wrong about just over a third of Democrats and just under a third of Republicans.
Myth 2: Statistical data obtained from rigorous surveys tell us how Americans really feel about abortion rights.
Although we can learn a lot from surveys, there are limits on how much statistics can reveal about attitudes toward complex topics (a theme that emerges throughout this newsletter). There's also a degree of laxness in how some journalists approach statistical data, as if the fact that something is expressed as a number means that it's objective and somehow immune to misrepresentation.
For example, consider again that Fox News article I mentioned in which the writer expressed a deeply pro-choice attitude. It's a well-written article. The arguments are clear and heartfelt. And yet, she repeatedly claimed that 80% of Americans want abortion to be safe, accessible, and legally unrestricted. For example:
"When the time comes to make this deeply personal choice, 80% of this country believes that you and your doctor should have the right to make it together, without the government holding you back."
I was baffled by this statistic, because it didn't align with any data I could recall. I did some digging and found a 2021 Gallup poll, administered to a nationally representative sample, in which people were asked the following question: "Do you think abortions should be legal under any circumstances, legal only under certain circumstances or illegal in all circumstances?" (The nature of the "circumstances" was not specified in the question.)
32% of respondents said abortion should be legal under any circumstances, while 48% said it should be legal only under certain circumstances. That's 80%. But those are two different groups of people. The writer's claim that 80% of Americans believe that women (and their doctors) should be able to choose without government "holding you back", is simply not supported by these data. In this survey, only 32% of Americans support unrestricted abortion. That's more consistent with the results of other surveys.
2. What can we say with confidence about public opinion on abortion rights?
It's easier to criticize than to create. In the first part of this newsletter, I attributed difficulties in understanding public opinion to problems with survey methodology and media coverage. That was relatively easy to do. In this part, I'll discuss how Americans "really" feel. Obviously, public opinion is mixed, and somewhat (but not consistently) divided along partisan lines. but what else can we say with some degree of confidence?
The way social scientists and pollsters usually address a question like this is to disaggregate their data, and one way of doing so is to report the results separately for each demographic group. For example, a nonpartisan Pew Research Center poll published this month showed that the percentages of people who think abortion should be legal in "most" or "all" cases is is greatest among women, adults under 30, Democrats, Asians, athiests, and those who know someone who's had an abortion. Other surveys add, for example, that support for legal abortion is greatest among those with college degrees, and those who live in certain states (mostly on the west or east coasts, with the greatest amount of support in Massachusetts, Vermont, and Washington DC).
Results can also be disaggregated by topic. For example, in the 2018 Gallup survey, when asked whether abortion should be legal if the woman's life would be endangered by carrying the baby to term, 83% said yes for the first trimester, as compared to 75% for the third trimester. However, if the woman had been the victim of rape or incest, 77% of respondents supported first trimester abortion, a figure that dropped to 52% for the third trimester.
Disaggregation is helpful, but it only takes us so far, because it sorts people and their beliefs into boxes. Boxes don't capture belief systems very well. In particular, when people are divided into too many demographic boxes, you learn nothing about why they believe what they do. A woman who completes the Gallup survey is not just a woman. She is also, say, an older Black protestant who leans Democratic and doesn't know anyone who's had an abortion. If you assume that each of these demographic labels matters, which one or ones most strongly influence her views on abortion rights? And, how do they influence her views?
An NCOR study (currently in press), using techniques other than disaggregation, shows that peoples’ views on specific topics tend to be correlated (presumably because these views are informed by broader perspectives on abortion rights). For instance, how favorably people view legal abortion when the mother's life is endangered is correlated with how favorably they view legal abortion when the mother had been the victim of rape or incest. (The numbers aren't identical, just moderately correlated.)
The NCOR paper also indicates relatively little polarization when results are combined across questions, and additional questions are added. In other surveys that pose the question, very few people, regardless of party affiliation, believe that abortion should be illegal in all cases, without exception. For example, according to a recent Pew survey, this is only 4% of Democrats and 13% of Republicans. In the NCOR study, even fewer people take this extreme position and hold that women forced to have unwanted babies should receive no financial or social support. On the other end of the spectrum, although 30% of Democrats (and leaners) and 6% of Republicans (and leaners) in the Pew survey support legally unrestricted access to abortion, fewer people take this extreme position without reservations in certain scenarios (e.g., pregnant adolescents).
Although some media reports may suggest otherwise, not that many people are simply pro-choice or pro-life in some easily characterized way. Adding to the survey data are narrative accounts indicating that many people have nuanced, conflicted, uncertain, and sometimes admittedly contradictory beliefs.
Summary
How can we characterize Americans public opinion on abortion rights? I mean, apart from saying that we're divided, tense, worried, conflicted, and uncertain?
The granularity of the survey data can be impressive, and yet, as I've said, they don't tell you everything. Narrative accounts shared by individuals are deeply moving, and remind us of the complexity of human experience, and yet in the end these are the stories of individuals rather than the entire group. So, maybe that's the summary statement? Maybe we should just say that we're a nation of individuals, each with our own take on these complex, interwoven, and painful issues?
I do see things that way. But I would add some generalizations. What follows is not a balanced summary of everything I've discussed previously, but rather what I take to be the most important points, including a few not mentioned earlier:
1. Gender is no longer a strong predictor of views on abortion rights. The Pew findings are typical, in that only 5% more women than men stated that abortion should be legal in most or all cases. Gender differences on more specific questions were comparably small. The largest of these differences was in the percentage of Democrats who say that women should have a greater role than men in setting U.S. abortion policy, but even that difference wasn't overwhelming (48% of women; 35% of men).
2. The strongest predictors of views on abortion rights tend to be political affiliation and religious beliefs, but these predictors are only strong in a relative sense (i.e., compared to other predictors). In an absolute sense, they're quite fallible. Generally speaking, knowing something about a person's political affiliation and demographics won’t give you a very reliable or precise understanding of their attitudes about abortion rights.
3. Attitudes toward specific abortion rights topics tend to be interrelated. For example, people who feel that abortion should be legal when the mother's life is endangered tend to also feel that abortion should be legal if the baby would be born with a life-threatening illness. However, attitudes toward specific scenarios are merely correlated rather than synonymous, and people sometimes contradict themselves.
4. Journalists, talking heads, and politicians tend to overestimate the extent to which people have clear opinions about legal access to abortion. Narrative accounts indicate that many of us feel conflicted and uncertain. In addition, not everyone fully understands the issues on which they express opinions. For example, not everyone understands what it would mean, legally speaking, for Roe v Wade to be repealed. Some studies suggest that as many as 15% of Americans cannot correctly articulate a single belief associated with either the Democratic or the Republican party (and thus they may inadvertently misrepresent themselves when indicating political affiliation on a survey). Another example is that some people who support a ban on abortion after 6 weeks believe that the heart literally starts beating at that time (actually, at 6 weeks the heart is just a tube of cells that emit synchronized electrical signals. At this point, the heart has no valves, no blood flow, and no audible "beat" except what’s created by a device that translates electrical signals into sound).
5. Very few Americans believe that abortion should be illegal, without exception, and that woman who are forced to have unwanted babies should receive zero support. Regardless of political affiliation, this is a fringe view.
6. The majority of Americans believe that abortion should be legal, at least under some circumstances. Many, but not all, of these people recognize that what women need is not just the legal right to have an abortion, but also legally protected access to safe and affordable procedures, as well as adequate postpartum support.
7. The majority of Americans want the Supreme Court to uphold Roe v Wade. People may not necessarily understand the legal details, but even a poll sponsored by Fox News this week reported that 63% of respondent want to "let it stand." (That's actually 2% higher than reported by Pew in March).
I'm not sure you and I can do anything that would substantially affect how the Supreme Court handles Roe v Wade, but you can certainly reach out to the two Senators representing your state and air your opinions for or against legislation that ensures legal protections pertaining to abortion. (Here's an article that lays out what the Women's Health Protection Act is, and why Congress voted it down yesterday 49-51.)