On the Turnaway Study

Speculations on why 38% of women change their minds a week after abortion denial, and other things

The Turnaway Study was a famous longitudinal study following women who received or were denied abortions in the U.S. from 2008 to 2010.

Most articles using data from the Turnaway Study focus on (a) the negative effects of abortion denial on health, happiness, and socioeconomic conditions, and (b) the fact that the vast majority of women who receive abortions are satisfied with this decision over both the short and long term. These findings are highly valuable and relevant to the conversation around reproductive health policy. As a broad stance, I think access to safe abortions is usually very important for improving women’s health, outcomes for families, preserving choice, etc.

How women perceive denied abortions five years later

One underanalysed area of the Turnaway data is how women who are denied abortions view their pregnancies afterwards. Occasionally, pro-life authors comment on the fact that, five years after being denied an abortion, 96% of women in the Turnaway study said they no longer wish they’d had the abortion. I’m unsure how much this can be seen as an indication of whether an abortion would have been the correct choice. Having a live four-year-old child probably emotionally biases one’s preference against having had an abortion. But it gets very confusing to define what a “correct choice” is in this case, and to distinguish emotional bias from a reasonable decision-making metric. How do we weigh the meaning/joy that a person derived from a child that would not have existed in another world?

My initial thought is that we can’t make decisions solely based on what we wouldn’t regret in the future. Firstly, because most women who received the abortion also didn’t regret their abortion, meaning either: (a) both options were equally good, or (b) humans subconsciously adjust their perception of past choices to justify their current situation. 

Secondly, regret-based decision-making leads us to all sorts of strange conclusions — we could make the argument that women should have as many children as biologically possible, because probably they won’t regret the existence of each individual child five years later when the child is a talking, walking, sentient creature. But it doesn’t seem desirable for women in most cases to have 17 children. This means the way we perceive a choice looking back, years later, may not be a perfect indication of how good that choice actually is for us right now. So the role of future regret in decisions is left in a tangled moral mess, and folds partially into a classic case of the non-identity problem. More on this some other time.

How women perceive denied abortions one week later

The much more interesting Turnaway statistic, in my opinion, is this: One week after abortion, "sixty-two percent of women in the turnaway group still wished that they had been able to obtain an abortion." This number is shockingly low — it means that 38% of women either completely change their minds or are unsure after a very short period of time.

There are a few readings of the situation. Firstly, it’s possible that the abortion would actually have been good for these women, but abortion denial causes some neural shenanigans and the women convince themselves it would’ve actually been bad. Choice-supportive bias, where people think their past decisions were more positive than they actually were, is supported by some studies. This bias happens through selective memory, misattribution, etc. 

But the bias only seems to be present where people have autonomy over their decisions, and thus want to believe that they had good judgment. Studies that focus on people’s perception of decisions that are assigned to them do not suggest a cognitive bias. This is a closer parallel to abortion denial, where women obviously do not have autonomy. The time frame is also significant here: is a week enough time for women’s brains to do a full 180 on a core life decision? This is unclear, and warrants more research.

A second explanation for our statistic is that these are women whose desire for an abortion may have been more shaky to begin with. The primary reason for abortion denial in the Turnaway samples was that women were too late into their pregnancy to legally receive abortions. Perhaps the group of women who request late term abortions were more indecisive earlier in their pregnancy, and thus more likely to also change their mind post-abortion denial. 

(As a caveat, indecisive women probably don’t represent the entire 38%. Lots of women in the study reported their partners being unsupportive of the abortion, or not having access to clinics nearby — these are other reasons why people might have requested later term abortions.)

Moral implications of the Turnaway study

The shaky motivations case is where I think some moral ambiguity arises. The argument most central to the pro-choice movement is that women ought to be able to make their own decisions, and this is laudable. Still, I do wonder if the consistency and intensity of someone’s preference towards abortion ought to be part of the equation. The 38% statistic indicates some subset of women would make a significantly different choice given more time or slightly different conditions. This might imply that their decision is less legitimate, depending on what we think legitimacy demands. In the 38% of cases, an instantaneous decision might have been misaligned with the person’s overarching values/goals, or misaligned with what the person themselves wanted at other points in time.

I don’t think there are clear-cut implications here for abortion policy. It would be helpful to have more qualitative data on what goes into the decision-making process of women who have voluntary, non medically/logistically necessitated late term abortions. This would illustrate where preferences change and how likely they are to be reflective of what a person generally wants at most points in time.

Mandatory waiting periods: a solution?

While taking into account the constraints of limited data, one possibility is that mandatory waiting periods (MWPs) for abortions might be more useful than the median progressive would believe. Especially in cases where the initial consultation can be performed virtually, a few days of waiting has a marginal harm to accessibility and potentially a large positive impact on people who would make life-altering choices differently. (I will caveat this by saying that where people have to consult in person and it is logistically unfeasible to travel long distances to a clinic/miss work multiple times/etc, I’m less certain that this tradeoff works.) 

Right now, MWPs are broadly frowned upon in reproductive healthcare. For example, the World Health Organisation argues MWPs “demean… women as competent decision-makers.” 

Firstly, if waiting periods are truly demeaning, I’m happy to demean all people as competent decision-makers and not just women. My framework of shifting human preferences isn’t unique to pregnant women in any way. Almost everyone has had the experience of making a decision at one point only to deeply regret it later. Many of the things that influence our decision-making — substances or illness, but also emotions or energy or who happens to be around us — are highly transient and yet affect us intensely. This tendency of the human mind to fluctuate and evolve is a core reason why I’m okay with the state intervening in people’s lives in limited cases. Laws aren’t intended just to protect others or preserve abstract moral values, but also to protect a future self who is in some sense morally distinct from the person right now.

But secondly, I don’t think it is demeaning. Traditionally, we view a person as an individual moral agent with a coherent, unified set of desires and obligations. We view morality as an exercise in balancing the preferences of agents against one another. For example, most discussion on abortion focuses on weighing women’s rights against the rights of unborn children. 

But sometimes, morality is also about balancing one person’s preferences against their other preferences. When our preferences change significantly over time, a person is sort of like a sequence of miniature moral agents who switch out for one another as psychological/environmental change occurs. Our treatment of Mini-Moral-Agent #1 that wants an abortion right now has an effect on Mini-Moral-Agent #2 who feels differently the week after. If a person’s instantaneous preferences are not necessarily representative of their longer-term preferences, honouring their autonomy sometimes means creating systems that allow for a more detailed and ongoing discussion of what the person wants.

This is reflected by lots of other social norms. MWPs are customary for medical assistance in dying in most countries in order to give people time to think and to ensure the sustainability of a decision. Lots of drugs are legally restricted, partially because we recognise a person’s immediate desire to get high at a party is not aligned with their future preference to live a long and healthy life. 

Richard’s thoughts on future preferences

Some authors disagree that future preferences are morally important. Richard Hanania had an interesting piece in his blog recently about why medical assistance in dying should be made more accessible. One point he makes is that a person’s instantaneous preferences are more relevant than what they might want in the future: 

“Opponents of euthanasia sometimes pretend as if they want people to make the decisions most consistent with their true beliefs, but it’s clear that they just want to stack the deck against them ending their own lives. For example, Raikin has talked about how some people have suicidal thoughts after say becoming severely disabled, but eventually learn to accept their condition. He argues that they therefore shouldn’t be given the option to end their lives when facing tough circumstances. Notice that he again treats the decision he likes as the correct one, and the decision he disapproves of as wrong. One could just as easily say that people who want to kill themselves after becoming disabled are seeing things more objectively, and when they accept their condition they’re coping and living under a kind of false consciousness. I’m pretty sure I would want to kill myself if I was ever paralyzed, and I grant it’s possible that maybe I would change my mind after a while. But I wouldn’t want to become reconciled to living with such a condition.” 

In the case of abortion, the argument might go like this: “A person who wants an abortion currently might change their mind, but that doesn’t mean the future decision is any more correct than the current one. And right now, they don’t want to change their mind.” 

This is an interesting argument, but not one that I think provides a particular justification for why our current preferences are more legitimate than future ones. My argument in the case of abortion is not that there is a universally correct decision on whether or not a person should have an abortion. It’s instead that future preferences are often more accurate. The advantage of your future self is that they can take into account a larger number of experiences and thoughts — the ones that were going on when you were inclined towards your original decision as well as everything that’s happened since then. A person a week after finding out they are pregnant can still remember the initial emotions they felt, but can also consider their reflections after thinking more or discussing with a partner or family. Unless we are given particular psychological reasoning to believe Hanania’s coping/false consciousness claim, I would usually defer to the later self. 

Decisions as weighty as abortion are complicated and individual, and creating systems that accommodate the wide array of circumstances is a thorny task. In some ways, the most interesting thing about the Turnaway data is the questions that it leaves unanswered. What exactly leads some women to regret and anger and others to reluctant acceptance after abortion denial? Why do some women land on abortion much later than others? What proportion of women would be meaningfully disadvantaged by mandatory waiting periods? A pro-choice movement that was less zealously committed to unconditional, instantaneous, individual decision-making might be more willing to find the answers.