Keeping secrets, both your own and others’, may seem like very personal business. However, it turns out that what sensitive information gets shared, and with whom, follows some clear social patterns. This is one of the big take-aways of Sarah Cowan’s study of how information about abortion and miscarriage circulates through social networks.
Cowan starts from the fact that even though abortion is a more frequent event in the U.S. than the miscarriage of a recognized pregnancy, “more Americans hear of women who have had miscarriages than they hear of women who have had abortions” (483). Using a nationally representative survey of 1,600 American adults, in fact, Cowan finds that each miscarriage “secret” was told to 2.63 people and kept from 0.2 people on average, whereas abortion secrets were told to only an average of 1.24 people and kept from 0.8 people.
Cowan suggests that the data show that abortion is a more stigmatizing than miscarriage (that is, it deviates further from social norms) as a piece of personal information. She cites higher levels of social disapproval and previous studies indicating that women frequently report their abortions as miscarriages to their doctors. In other words, stigmatized or potentially stigmatized information is less likely to be shared with others.
In addition, Cowan finds that secret telling/keeping is impacted by the presumed attitudes of its potential recipient. In this case, respondents who have more accepting views toward abortion are more likely to hear others’ abortion secrets. Controlling for how likely one is to hear secrets, for example, Cowan shows that staunch “pro-life” Americans are 58% less likely than are those who think abortion should be “generally available” to hear an abortion secret.
Cowan’s results highlight how selective information sharing and secret telling is, and that people are often only told of secrets with content they already approve of. This selective information flow can lead us to perceive that our social networks match our beliefs at a greater extent than they actually do.