Do women who drop out later regret their choices? Robin Abcarian writes today in the LA Times about a study that concludes:
“Tons of studies have shown mixed feelings of relief to be the most common short-term emotion after an abortion,” said the paper’s lead author, UCSF epidemiologist Corinne Rocca. “But even after five years, when emotions are pretty low, relief is still the most common.” She found that about 99% of women who had abortions told researchers five years later that they had made the right decision.
…Of course, some women have abandoned the study over the five years. Critics indicated that they were women who felt regret, shame or guilt about abortions. But Rocca analyzed exactly who got out and found no correlation between dropouts and negative emotions. “People can say that,” she said, “but I haven’t found any of it.”
This is . , , not entirely honest. Of course, people dropped out. People always stop longitudinal studies like this. The real criticism of Rocca’s study is that it only included women who agreed to participate in the first place. In other words, your sample is self-selected and not random. She defends this:
38% of the participants in a five-year study in which women are asked about a stigmatized health service take part in other large prospective studies. With the exception that the women in this sample were poorer, they were demographically similar to US women with unwanted pregnancies. Women showed a number of emotions when they enrolled: about two thirds expressed grief and over a third regretted it. We have no reason to believe that women will be included in the study based on how these emotions would develop over three years.
It is true that longitudinal studies by definition only include people who have first consented to participate in the study. If you’re studying a specific physical phenomenon like lead poisoning or the effects of a new drug, it’s probably fine. For this very reason, longitudinal studies are generally not suitable for assessing emotional states that can easily have an unexpected impact on participation. Rocca says, “we have no reason to believe” that this has happened here, but that’s a rather unconventional approach to legitimate criticism. I can think of half a dozen reasons why emotional states can influence participation. Maybe people who feel guilty want to be reminded less of it over the next five years. Perhaps introverts are less inclined to participate. Perhaps women with traditional upbringing participate less. This is an endless list, and “no reason to believe” usually indicates that no one has bothered to look for it.
None of this means that the study’s conclusions are wrong. I assume that this is basically correct. However, I don’t understand how you can claim to get reliable results from an extremely non-random sample like this one. This applies to both studies that we agree with and studies that we disagree with.
¹ It is worth noting that this also applies to survey-based studies. This is only a basic problem when dealing with emotional abortion reactions. Many people are unlikely to want to participate, and it is next to impossible to know for certain whether this affects the randomness of your sample.