Dobbs' Impact on Women
My newsletters often focus on statistical errors and deceptions. This week, I'll be showing how statistics can provide uniquely powerful, nimble ways of tracking the influence of new legislation.
A methodological challenge
On June 24, the U.S. Supreme Court ruled in Dobbs v. Jackson that our Constitution does not guarantee the right to an abortion.
Legally speaking, the impact of Dobbs is complicated, because the 50 states embody 50 different sets of abortion laws, many of which changed following Dobbs, or are in the process of changing. Most of the changes consist of greater restrictions or bans. (Only 10 states have passed new laws guaranteeing a legal right to abortion and/or greater access; this includes ballot initiatives passed on Tuesday.)
How do the new legal restrictions affect women who seek an abortion? Are some groups of women more adversely affected than others? These are important questions, but we might wonder how they can be addressed given that, legally speaking, each state tends to be unique. As the laws rapidly evolve, we might ask too how researchers can keep up.
Meeting the challenge
Studies on the impact of Dobbs are beginning to appear, including four published this November 1 in a special issue of the Journal of the American Medical Association (JAMA). These studies are enabled by technologies that support big data methods and data sharing, as well as statistical techniques that allow researchers to analyze data from states with diverse legal environments.
I'll be discussing some of the new studies here, interleaved with older ones. I'll start with changes in abortion rates, then look at the connection between Dobbs and women's well-being. Much of this is grim stuff, but at the end, in the Conclusion, I'll share a bit of good news.
Full disclosure: This newsletter is influenced by a liberal and strongly pro-choice perspective. I wouldn't have it any other way. But the new studies are relevant to everyone, regardless of political perspective. The data are what they are.
Clinical abortion rates
A 2017 study estimated that roughly 1 in 4 American women will seek an abortion in her lifetime. However, the number of women who actually do so each year is uncertain. For the three years prior to the pandemic, for example, estimates range from just over 600,000 per year (CDC) to over 900,000 per year (Guttmacher Institute). The latter is probably more accurate (only 47 states report data to the CDC) and yet may still be an underestimate, given that reporting is voluntary, and researchers can only guess at the rates of both legal and illegal self-managed abortions.
Although we're uncertain about absolute abortion rates, the impact of Dobbs on changes in those rates seems a little clearer. This is illustrated by an October 28 report from the Society of Family Planning (SFP).
SFP analyzed state-level data from abortion care providers for April through August 2022. (Dobbs was decided on June 24.) Each state was classified as either banned, restricted, or legal, based on the nature of their abortion laws following Dobbs. Here are some changes in abortion rates observed between April and August:
—States where banned: 95% decline
—States where restricted: 32% decline
—States where legal: 11% increase
—All states combined: 6% decline
These data suggest that following Dobbs, many women living in states where abortion became restricted or banned traveled to states with more liberal policies to get an abortion, and yet, overall, fewer women sought abortions. (Evidence from other years confirms that the overall 6% decline for this time period is atypical.)
The same pattern was observed in a separate study focusing specifically on Texas. Following the passage of a near-total ban on abortion in September 2021, Texas woman obtained more out-of-state abortions and fewer ones in-state, but the overall rate declined. Still other studies show that in Texas, and nationwide, poor women and women of color (particularly Black women and Latinas) experience the greatest financial- and transportation-related deterrents to clinical abortion, and thus have higher rates of unwanted pregnancies (a problem that in turn contributes to higher rates of maternal mortality among these groups).
Overall declines in abortions following Dobbs are unsurprising, given, among other things, the logistical challenges of traveling across state lines for a medical procedure. As you can see on the Center for Reproductive Rights map below, more than half of U.S. states are classified as "not protected", "hostile", or "illegal" concerning abortion law, and these states tend to cluster. This isn't meant to downplay the burden of within-state travel, which is prohibitive for some women, but to emphasize disparities in the extent of cross-state travel that may be needed. Experts have noted that abortion laws are most restrictive in the poorest states, most of which are located in the Southeast.
The data I've discussed so far pertain to clinical abortions. What about those that are self-managed, by means of prescribed medication? Is it possible that overall abortion rates have remained unchanged, because the women who desire them but can't travel to a clinical setting are resorting to pills instead?
Self-managed abortion rates
We know at least that requests for abortion pills are increasing. In another November 1 study, researchers looked at data from Aid Access, the only organization that provides both online telemedicine service and medication for self-managed abortions. The focus was on the 30 states where Aid Access is permitted to operate completely separately from a formal health care setting such as a hospital or clinic.
Requests were examined for three time periods: before the Dobbs decision was leaked (September 1 – May 1, 2021), between the leak and the formal announcement (May 1, 2021 – June 23, 2022), and after the formal announcement (June 24 – August 31, 2022). The average number of daily requests for abortion pills increased from 82.6 prior to the leak, to 137.1 after the leak, and 213.7 following the announcement, a significant difference in each case. Increases were observed in every state, but the relative changes were greatest for states with total bans on abortions.
Given that self-managed medical abortions are increasing, we might ask: Does this increase match the decline in clinical abortions? In other words, are the overall rates remaining constant as the methods change?
Most probably not. State-level abortion bans apply to medication as well, and 19 states currently require prescribers to be physically present when the medication is taken. Put simply, access to pills is not always easier than a clinical approach if one is playing by the rules.
There's also data on this issue. The Upshot – a website published by the New York Times (and, incidentally, one of the most reliable sources of data journalism out there) – combined SFP's finding of a 6% overall decline in clinical abortions with the increase in abortion pill orders and estimated that, nationally, abortion rates via any legal method have decreased 2% since Dobbs.
One strength of The Upshot analysis is that abortion pill orders were considered from other overseas sources besides Aid Access. Even so, uncertainties remain. Data were not available from all overseas providers. There's no way of knowing how many women elected not to take the medicine (or decided to seek clinical care). So, even though national abortion rates have declined following Dobbs, current estimates may be wrong about the extent of decline.
In sum, the new data illustrate a cascade of events leading from Dobbs to more restrictive state laws to lower abortion rates, particularly among poor women and women of color, a decline that has not been offset by access to abortion pills. These findings raise questions about the well-being of women who want an abortion but, for whatever reason, are unable to obtain one. Such women have always existed, but there are more of them now following Dobbs.
The consequences of unintended pregnancy
If more restrictive abortion laws are resulting in fewer abortions, then more unintended pregnancies are presumably being carried to term. What impact does this have on women (and their children)?
Between 2017 and 2019, roughly 38% of births in the U.S. resulted from unintended pregnancies (i.e., pregnancies that the mothers reported not wanting, at least at the time they discovered they were pregnant.) In a November 1 study, researchers published the first systematic meta-analysis concerning the impact of these pregnancies on infant/maternal outcomes. This meta-analysis considered 36 studies with a total of 524,522 participants. Criteria for inclusion were stringent, meaning that only methodologically strong studies were analyzed. The main findings were that prenatal depression, postpartum depression, maternal experience of violence, preterm birth, and low birth weight are each significantly worse among woman who describe their pregnancies as unintended.
These findings add to the bad news we've learned from earlier research, such as the Turnaway Study, a five-year longitudinal project that tracked the impact of unintended pregnancies that occurred between 2008 and 2010.
A distinctive feature of the Turnaway Study is that the roughly 1,000 participants included a group of women who received an abortion because they came to a clinic within two weeks of the legally-imposed gestational limit, as well as a group of women who were denied an abortion because they came to the clinic up to three weeks past that limit. Since both groups of women wanted abortions – and sought one at roughly the same point in their pregnancies – the key difference between the groups was simply whether or not they received one. This is stronger, methodologically, than a study that simply compares women who do vs. do not have the procedure. (Other methods were used to ensure that the groups were comparable. A third group received an abortion during their first trimester.)
Turnaway Study data has fostered more than 50 separate peer-reviewed studies. Here are a few of the many findings:
(a) Compared to women who received an abortion, those who were denied one experience higher rates of poverty lasting at least four years, as well as other material hardships and poorer health.
(b) Women denied an abortion are more likely to remain with an abusive partner and continue to experience abuse. (Those who receive an abortion experience subsequent declines in intimate partner violence.)
(c) The babies of women denied an abortion show slower development in a variety of domains for at least five years.
Other studies have shown that most of these effects are, or are expected to be, worse among poor women, as well as Black women and Latinas.
Conclusions and good news
At least five new studies published this week are consistent with, and extend, the findings of prior research. These studies tell us that Dobbs v. Jackson allowed states to pass more restrictive abortion laws, and these laws have in turn diminished overall abortion rates. Access to medication for self-managed abortions has partly but not completely offset the decline in clinical abortions. The net result is more unintended pregnancies, which are linked to a variety of negative outcomes for mothers and their children, and are especially problematic for poor women and women of color.
The new data remind us that apart from one's political views, it's clear that restricting access to abortions creates harm for women. In my view, this is a key part of the rationale for a pro-choice view (along with a woman's right to control her own body, evidence that most Americans are now pro-choice, and so on).
Women are most directly and negatively impacted by restrictive abortion laws – the extent of harm is deeper and more complex than I can do justice to here – but other groups are affected too, including women's families, the babies they carry to term, and the medical providers, pharmacists, and even advertisers who incur legal risks by supporting access to abortion.
Creating laws that are more supportive of women's reproductive rights will continue to be challenging, because people on all sides of the debate tend to be passionately committed to their beliefs. I want to close with a statistical perspective on this issue, accompanied by a bit of good news. (Or, if you happen to be pro-life, bad news.)
The country as a whole leans pro-choice, but, as we know, views on abortion rights are linked to political affiliation. In a March 2022 Pew Research poll, for instance, more than twice as many Democrats as Republicans stated that abortion should be legal in "all or most cases". That seems like a big difference, but the actual numbers were 80% (Democrats) versus 38% (Republicans), meaning that the latter aren't far from a 50-50 split. Since 2007, the percentages of Republicans who give "all or most case" responses to this question have remained remarkably constant.
Here's the good news: More than half of Republicans under 30 are providing such responses. It's older Republicans who are increasingly responsible for the party's overall opposition to abortion. Within both parties (and Independents) there's a tendency for younger people to view abortion rights more favorably than their parents do.
This is good news in the long run. Although it's no consolation to a woman in Texas right now who lacks time, strength, or resources to travel hundreds of miles to get an abortion, it does suggest a better future.
In that hopefully not-so-distant future, will younger people actually turn out to vote for pro-choice candidates and initiatives? The midterms are encouraging in this regard – the turnout rate for people under 30 was 27%, which is low in an absolute sense but still the second highest rate for a midterm in the past 30 years. And so, I remain hopeful.
Thanks for reading!