Summary

Researchers tested 85 gender-nonconforming young children, none yet socially transitioned, on tasks about gender identity and liking for toys, clothes and playmates. About two years later they checked whether the children had transitioned (changed pronouns). 36 had. Children who identified more strongly with the other sex at first testing were more likely to have transitioned, and their scores already looked like those of transitioned children and of children of the other sex. The authors conclude that childhood transitions are not random, and that transition itself does not seem to strengthen identification with the other sex. But the group was small, self-selected and from liberal, well-off families. Follow-up was short. The result was a parents' decision, not long-term identity. And the 'no change' claim compares different children, not the same child before and after.

Key findings

• 85 kjønnsavvikende barn som ikke hadde gjennomgått sosial overgang ble testet og fulgt opp omtrent 25 måneder senere i gjennomsnitt.
• 36 barn (41 %) hadde gjennomgått sosial overgang ved å endre pronomen ved oppfølgingen. 49 hadde ikke.
• Barn som identifiserte seg sterkere med det andre kjønnet ved starten hadde omtrent 4 ganger så stor sannsynlighet for å gå gjennom overgang (omtrent 5 ganger etter å ha tatt hensyn til andre faktorer).
• Et barn med en middels skår hadde omtrent 30 % predikert sjanse for å gå gjennom overgang, mot omtrent 48 % for et barn med høy skår.
• Barn som gikk videre til overgang skåret omtrent likt som barn som allerede hadde gjennomgått overgang og som barn av det andre kjønnet.
• Gutter hadde omtrent 4 ganger så stor sannsynlighet for å gå gjennom overgang som jenter, og et lengre tidsrom før oppfølgingen forutså også overgang.
• Foreldrene var svært politisk liberale og høyinntektsfamilier, og forfatterne sier at overgangsraten på 41 % trolig overdriver raten for kjønnsavvikende barn generelt.

Abstract

Increasing numbers of gender-nonconforming children are socially transitioning-changing pronouns to live as their identified genders. We studied a cohort of gender-nonconforming children (n = 85) and contacted them again approximately 2 years later. When recontacted, 36 of the children had socially transitioned. We found that stronger cross-sex identification and preferences expressed by gender-nonconforming children at initial testing predicted whether they later socially transitioned. We then compared the gender-nonconforming children with groups of transitioned transgender children (n = 84) and gender-conforming controls (n = 85). Children from our longitudinal cohort who would later transition were highly similar to transgender children (children who had already socially transitioned) and to control children of the gender to which they would eventually transition. Gender-nonconforming children who would not go on to transition were different from these groups. These results suggest that (a) social transitions may be predictable from gender identification and preferences and (b) gender identification and preferences may not meaningfully differ before and after social transitions.

Conclusion

Despite limitations and a need for future replications, this study provides preliminary evidence that extremity of identification with the gender opposite one's assigned sex predicts childhood social transitions. Moreover, differences in gender extremity likely exist prior to—and not because of—social transitions.