Skip to main content
ABC News
It’s Really Hard To Measure The Effects Of Abortion Restrictions In Texas

Last summer, after Wendy Davis had come and gone, the Texas legislature passed a package of abortion bills that has effectively forced most of the state’s clinics to close. The bills didn’t ban abortion outright, but instead placed new restrictions on abortion providers, such as a mandate for expensive structural changes (e.g. wider hallways and new ventilation systems) for clinics. Proponents said that more rigorous standards would protect women’s health, but obstetricians and pro-choice advocates warned that the law would serve a pro-life agenda, and that its consequences could radically alter options for women in the state. Left with only a handful of clinics in large cities at the center of Texas, women outside urban areas might take matters into their own hands and begin inducing abortions themselves.

Nineteen of the state’s 41 clinics have already shut down after the first wave of tough new standards went into effect last November. Come Sept. 1, 16 more are slated to close, leaving six abortion providers for a state of nearly 5.5 million reproductive-age women. After the first wave of closures, researchers at the University of Texas’s Policy Evaluation Project surveyed the state’s remaining abortion providers and reported that the abortion rate (the number of abortions per 1,000 women ages 15 to 44) had already fallen by 13 percent compared to the previous year.

But what’s behind that 13 percent number? Are women making illicit attempts to end their pregnancies, carrying them to term, or leaving the state to find an abortion elsewhere? To what extent is the abortion rate being influenced by factors that have nothing to do with the clinic closures? The answer is more complicated than the political debate suggests. As abortion becomes more difficult to access, so does abortion data.

The specter of self-attempted abortion conjures up the days before abortion was legal, when thousands of women were admitted to hospitals each year with hemorrhaging and infections because of incomplete self-induced abortions. According to one researcher I spoke with, women in the Rio Grande Valley, a slice of floodplain at Texas’s southernmost tip, may already be familiar with self-induction methods, which are common across the border in Mexico. The region, which is home to some of the most impoverished counties in the nation, used to have two abortion clinics. But they closed earlier this year, and now the nearest abortion provider is a four-hour drive away.

“One important reason that women turn to self-induction is because of a lack of clinic-based care,” said Dan Grossman, a co-investigator of the Texas Policy Evaluation Project. “It’s a hypothesis, but it seems likely that given the clinic closures, greater knowledge about self-induction methods, and the high rates of poverty in the area, that this is something more women are going to consider.”

Most of this, as Grossman acknowledges, is conjecture. Even if the restrictions on clinics do result in an uptick in women’s attempts to end their pregnancies on their own, the trend will be virtually impossible to identify. “Self-induced abortion is kind of invisible,” said Rachel Jones, a senior fellow at the Guttmacher Institute, a reproductive health research organization. “It’s by definition unobservable. If a woman is successful, we don’t know about it, so we can’t count it.”

Statistics on the observed abortion rate are difficult enough to gather. The CDC publishes abortion data culled from state health departments, but these numbers are incomplete. At least five states — including California, home to one-eighth of the U.S. population — simply don’t count the number of abortions performed within their borders. To fill this hole in the data, the Guttmacher Institute performs an abortion “census” every three years or so by surveying every abortion provider in the country. Jones has helped to conduct Guttmacher’s census for a decade and said that as more states pass laws like the one in Texas, the survey becomes more difficult to execute. “Abortion providers are under more scrutiny these days, so they’re sensitive about releasing this kind of information,” she said. “And even if they want to, they’re busier than they used to be. It’s hard to find time to fill out our form.”

Previous attempts to pin down the number of black-market abortions have only captured a small subset of women who might self-induce. In 2008, Guttmacher conducted a nationally representative survey of patients at abortion clinics and found that only 1.2 percent of women seeking an abortion had used misoprostol, an ulcer medication that can induce a miscarriage, to “bring back” their periods or to end their pregnancies. A similar number, 1.4 percent, said they had used other substances — like herbs or Vitamin C — to self-induce. Jones, who was a lead author on the study, said the methodology was not “optimal” because it failed to capture women who had attempted to self-induce but didn’t seek clinic care. Guttmacher isn’t currently planning another attempt to quantify self-induced abortion, she said, because at the national level, “it’s too overwhelming to estimate.”

Self-induction isn’t the only remaining way for women in Texas to have an abortion. New Mexico, Oklahoma, Louisiana and Arkansas have seven clinics within 200 miles of the Texas border — including clinics in Shreveport, Louisiana, which is 22 miles over the eastern border, and Santa Teresa, New Mexico, which is 15 miles over the western border. Past experience suggests women will seek out their services.

Ted Joyce, an economics professor at the City University of New York, is one of the few researchers who has used data to assess whether state-level abortion regulations have resulted in fewer abortions. In the 10 years since he began to evaluate the effects of abortion policy, the political debate over restrictions on clinics has become increasingly contentious, and states have responded by clamping down on scholars’ access to the data. But even before the numbers were politicized, drawing conclusions about individual laws’ effects was no easy feat.

Joyce discovered as much when he studied the effects of an earlier set of abortion restrictions in Texas. In 2004, Texas began enforcement of the Women’s Right to Know Act, a law that — in addition to mandating that women seeking abortions receive information about alternatives — effectively banned abortions after 16 weeks of gestation. To determine whether the law would decrease demand for abortions later in pregnancy, Joyce first turned to Texas’s abortion data, and found that the number of abortions performed at or after 16 weeks plummeted by 88 percent after the law went into effect, from 3,642 in 2003 to 446 in 2004.

It seemed like the law was working, but it wasn’t clear whether women were carrying their pregnancies to term, or finding other ways to end them. Joyce examined abortion statistics in neighboring states and tabulated the patients’ states of residence, looking to see how many came from Texas.1

The number of out-of-state abortions to Texas residents at or after 16 weeks’ gestation increased by nearly 300 percent, from 187 abortions in 2003 to 736 in 2004, indicating that many women were obtaining the procedure across state lines.

But the rise in out-of-state abortions didn’t account for all — or even most — of Texas’s decline. Joyce estimated that even with the 300 percent increase in out-of-state procedures, Texas’s abortion rate at or after 16 weeks still fell by 69 percent after the Women’s Right to Know Act went into effect.

It was still risky, though, to assume that those women were resigning themselves to their pregnancies, because they may simply have been getting abortions before 16 weeks. So that set Joyce on a different path: to somehow find out whether there was an uptick in births, and to look at whether there were changes in abortion rates earlier in pregnancies. Both questions came with their own issues.

Waiting for the state’s birth rate to rise is, according to Joyce, a generally ineffective strategy. “If the abortion rate declines by 13 percent, you’re talking about maybe 10,000 extra births in a state where 400,000 babies are born every year,” he said. “That 10,000-birth increase could have to do with migration, or the economy, or any number of things. You’d have to see a huge change before you could attribute it to abortion.” In the end, Joyce concluded that because the abortion rate before 16 weeks stayed steady, most of the decline in second-trimester abortions caused by the Women’s Right to Know Act resulted in births.

This last strategy — looking at abortion rates earlier in pregnancy — is less applicable after the most recent laws, because the new restrictions on clinics will make abortion less accessible throughout pregnancy. A limit on abortion after 16 weeks is less far-reaching than a reduction in supply across the board, since the vast majority of abortions take place before 12 weeks. And Joyce’s workarounds only narrow down the number of women who aren’t getting legal abortions and don’t necessarily indicate a spike in black-market terminations.

Even the 13 percent drop in Texas’s abortion rate over the past year is, Joyce said, overstated. Like the abortion rate nationwide, the number of abortions performed in Texas has been declining by 5 percent to 6 percent per year, independent of policy changes. That means only 7 percent to 8 percent of the drop can be attributed to the new law. “You can imagine, based on the previous research, that a certain proportion of those women are going out of state, so the actual decline in abortions for Texas residents is even lower,” Joyce said. “Is it possible that some of those remaining women are self-inducing? Is it possible that they’re continuing with their pregnancy? We don’t know, and the birth data can’t tell us because we’re talking about such a small number of potential births.”

Right now, sticking to surveys seems like the most feasible approach for researchers who are trying to measure self-induced abortion. Grossman said that the Texas Policy Evaluation Project has plans to survey a representative sample of reproductive-age Texas women to gauge their familiarity with self-induction methods and see whether they know someone who has ended a pregnancy on her own. They also hope, at some point in the future, to gather abortion data from Texas’s neighbors to determine how many women seeking abortions are being diverted out of state.

Even this, he admits, isn’t an ideal method. But Grossman’s goal is broader than pinning down the number of self-induced abortions — he wants to ensure that the outcomes of the state’s policies are studied at all. It’s a different type of struggle, largely because proponents of abortion restrictions like those in Texas seem uninterested in evaluating their effects. “These restrictions were passed based on no evidence, and they’re being defended based on no evidence,” he said.

But any proof that self-induced abortion is becoming more widespread will likely remain an anecdote rather than a data point. “There are so many reasons why an abortion rate might decline, and self-induction is the hardest to measure,” Joyce said. “I’m not saying it can’t be done, but it’s going to be seriously difficult to prove.”


  1. Analyzing the extent of this trend comes with its own challenges, however, not least of which is the task of corralling intricate data from five or six health departments. For his study, Joyce gathered abortion data from Arkansas, Kansas, New Mexico and Oklahoma from 2001 to 2006. Statistics from Louisiana, Texas’s neighbor to the east, were harder to come by, in part because residency status wasn’t reported on abortion certificates in the state until 2004.

Amelia Thomson-DeVeaux is a senior editor and senior reporter for FiveThirtyEight.