August 9, 2017

A recent article appearing in the journal Pediatrics has gone where few studies before it have trekked, assessing the mental health of 73 transgender children ages 3-12 whose parents have supported their social transition. That is, these are children who were observed as one sex when they were born but—according to their parents—are presenting themselves to their peers and families as the opposite sex. It’s a social transition—no radical surgeries or hormonal treatments have been undertaken yet.

The study’s authors found no differences among socially transgendered children and their siblings—as well as a matched control group—on depression, but slightly elevated levels of anxiety. The media called it “a first-of-its-kind longitudinal analysis of the mental health of transgender children.” Critics, they claim, are framing the parents’ decisions to support their children’s transition as flippant or rushed, or irreversible—all of which are untrue, the authors assert. I agree. I suspect few parents are in a hurry here, and a social transition is not nearly so difficult to reverse as hormones and surgeries are.

And yet this is not the study from which any clear protocol could sensibly be realized. Why? Here are three reasons:

First, there is a low ceiling to what we can know with confidence here. Small samples preclude much statistical power to detect real differences that may (or may not) actually exist. When no statistical differences are detected in popular domains like this one, it leads enthusiastic journalists—and complicit authors—to conclude that everyone is fine, when the social reality may be otherwise. All the qualifiers and limitations, which are important, are buried at the end of the article—as is standard—but they are many and wind up reading like the monotone final half of a prescription drug commercial. They note selectivity concerns that these may be parents who make their children feel “supported in general,” not about gender identity in particular.

Second, the authors admit concerns that “parents of transgender children could have biased reporting, reflecting a desire for their children to appear healthier than they are.” They’re right. A small number of parents known to be invested in their children’s transitions, recruited from transgender support groups, may not be stably valid, unbiased sources of information about their children. This is what experimental researchers call the “Hawthorne effect.” It occurs when persons being studied alter their behavior as a result of their awareness of being studied. If I was such a parent, I would want the data to reinforce the social support I was giving my child. And as that parent, I would have the opportunity to offer data in keeping with that social support. It’s not just theoretical. The Hawthorne effect is still a problem. All the more, I hold, when the study matter concerns politically sensitive matters. As the authors aptly note, a teacher’s perspective can be rather helpful here. They are outsiders, so to speak, at much lower risk of the Hawthorne effect than parents. Conceivably, too, instructors don’t or shouldn’t have a great deal invested in what the best course for navigating a particular child’s gender dysphoria ought to be. Sampled in larger numbers, a teacher’s perspective is an optimal route, especially in assessments of external behavior.

Third, the authors’ key interest is in discerning “internalizing psychopathology,” rather than documenting externalizing behavior. In other words, they are asking motivated parents to tell us what’s going on inside a four-year-old’s mind, rather than telling us what they observe in their children’s words and actions.

Fourth, “prejudice,” “discrimination,” “stigma,” and “rejection” are unmeasured specters in this study. They are blamed, as usual, for any negative experiences transgender children, adolescents, or adults report in more numerous previous studies, as well as the higher anxiety reported in this one: “(d)espite receiving considerable support from their families, these children likely still experience relatively high rates of peer victimization or smaller daily micro-aggressions,” particularly from peers. I wouldn’t dream of suggesting these forces do not exist, or are not harmful. What is lamentable is that few studies include available measures of them so scholars might actually assess their roles in understanding the pathways between transgender decisions and mental health challenges.

Other questions arise, further taxing an already small sample. How long had the children been socially transitioned? Since before kindergarten, or well after they had begun their schooling? Had they changed schools? In other words, how many of their classmates knew them prior to their social transition? Would socially transitioning back to their natal sex create temporary emotional challenges? (I can imagine it may.)

I’m not the study’s only critic. Indeed, the authors themselves note that they are swimming upstream: “By and large, studies of children with GID (gender identity disorder) reported high rates of psychopathology, especially internalizing disorders” while only two smaller studies suggest “relatively good mental health.” That pair of studies has “received some criticism for methodologic limitations” and diminutive sample sizes. In other words, the authors acknowledge their study is unusual in its surprising results. Moreover, in a follow-up letter to Pediatrics, three Mayo Clinic child and adolescent psychiatrists expressed “concerns about the meaning and generalizability of their findings.” In particular, the study’s scales were not intended for use among the 30 percent of children in the study who were between 3 and 5 years old. Moreover, they noted, the sample is biased in other ways besides those already mentioned; it is overrepresented by higher-SES parents, the kind of parents that are more responsive to problems and whose children tend to exhibit fewer internal and externalizing problems in the first place. The high-SES problem sounds familiar; the social science debate about gay parenting was dominated by a focus on privileged households.

I am a big fan of data collection on sensitive domains such as this one, and of federal funding of the same. Were we to collect more high quality, multi-source data from larger—and more random—samples of such children (and at older ages), the medical and social scientific community would be in a far better position to assess the wisdom of emerging protocols and peer pressures in this domain.

About the Author

Mark Regnerus is an associate professor of sociology at the University of Texas at Austin and a senior fellow at the Austin Institute for the Study of Family and Culture. His research is in the areas of sexual behavior, family, marriage, and religion. Mark is the author of over 40 published articles and book chapters, and three books.

About our Editors