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

• 社会的トランジションをしていない、性別に非同調的な子ども85人がテストを受け、平均約25か月後に追跡されました。
• 追跡時に、36人の子ども（41%）が代名詞を変えて社会的トランジションをしており、49人はしていませんでした。
• 最初に異性をより強く自認していた子どもは、トランジションする可能性が約4倍高くなっていました（他の要因を考慮すると約5倍）。
• 中程度のスコアの子どもがトランジションする予測確率は約30%、高いスコアの子どもでは約48%でした。
• その後トランジションした子どもは、すでにトランジションしていた子どもや異性の子どもとほぼ同じスコアでした。
• 男児は女児の約4倍トランジションしやすく、追跡までの期間が長いこともトランジションを予測しました。
• 親は政治的に非常にリベラルで、収入が高く、著者らは、41%というトランジションの割合は、性別に非同調的な子ども全般の割合を過大評価している可能性が高いと述べています。

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.