One of the least helpful habits in discussions of gender dysphoria is treating a shared description of distress as evidence of a shared developmental history. Once everyone has been placed under the same heading, the differences that might matter clinically become surprisingly easy to overlook.
A new paper led by Hollie Hammond, a PhD and Master of Clinical Psychology candidate, brings those differences into focus. I co-authored the study with Hollie, Emma Walter, Martin Daly and Evelyn Smith. It was published in Archives of Sexual Behavior on 19 September 2026.
We examined gender dysphoria in 138 adolescents and young adults aged 16–25 who were female at birth, often described as AFAB. Participants were recruited online and through word of mouth, and reported their own experiences. The question was whether different recalled childhood histories were associated with different current clinical and demographic features.
What we found
Using retrospective responses to adapted childhood gender-dysphoria symptom criteria, 29 participants were classified as early-onset and 109 as late-onset: 21% and 79% of this sample.
The early-onset group reported more severe current gender dysphoria, higher anxiety and a greater number of professionally diagnosed comorbid conditions. The late-onset group was more likely to endorse a nonbinary identity. Depression, stress and well-being did not differ significantly between the groups.
That pattern is worth pausing over. An account that simply assumes later-onset presentations must carry the greater psychiatric burden would have missed the direction of these findings. Equally, lower average anxiety or dysphoria in the late-onset group does not mean this group was free of distress.
The results make heterogeneity harder to ignore. They also make a convenient clinical story harder to assume.
What “late-onset” means here
The classification needs care. Participants recalled symptoms between ages three and eleven. Those reporting at least six childhood symptoms lasting at least six months were classified as early-onset; those below that symptom threshold were classified as late-onset. This was a retrospective research classification, rather than a childhood diagnosis established through contemporaneous assessment.
Consequently, “late-onset” does not mean every participant had no childhood gender nonconformity or no earlier gender-related feelings. The group recalled fewer than two childhood symptoms on average. That is informative, but it is different from claiming a universal absence of childhood signs.
Nor should the 79% figure be read as a population prevalence estimate. This was an uneven convenience sample, with 29 people in one group and 109 in the other. Memories of childhood may be influenced by present experience, and the study itself discusses whether its early-onset group fully resembles the very young, prospectively observed clinical cohorts described in earlier research.
Even the childhood-symptom difference needs interpretation: the groups were created using those symptoms. Their separation on that measure is partly built into the classification. The associations with other features are where the comparison becomes more informative.
A shared label leaves important questions unanswered
The clinical question I take from this work is how to understand the particular developmental history in front of us. When did discomfort begin? What form did it take? How did puberty, bodily changes, relationships, sexuality, mental health and the person’s developing self-understanding figure in that history?
Those questions should be available to investigation. A person's current identity cannot answer all of them retrospectively.
There is also a distinction between identifying developmental differences and establishing their causes. This cross-sectional study cannot establish whether dysphoria preceded other psychological difficulties, whether those difficulties contributed to dysphoria, or whether other factors shaped both.
It also does not demonstrate that late-onset gender dysphoria is socially caused. The groups did not differ significantly in the reported number of transgender or gender-diverse friends before participants recognised their own dysphoria. That finding belongs in the account, including when it complicates an attractive hypothesis.
Likewise, the study did not test which treatment works best for which group. Developmental heterogeneity gives us a reason to investigate that question more carefully; it does not supply a treatment algorithm.
Let the differences matter
I would like a field in which it is unremarkable to ask whether people describing similar distress arrived there through different routes. We routinely need that question elsewhere in clinical psychology. Gender dysphoria warrants the same curiosity.
Hollie’s work contributes to that task. It identifies differences worth pursuing while leaving the causal and longitudinal questions open to properly designed research. The next step is to follow developmental trajectories and outcomes closely enough that our explanations can be tested.
Good clinical science must be able to discover that its categories contain more variation than its preferred story allows.
Read the open-access paper or my original LinkedIn post.
Hammond, H., Walter, E., Daly, M., Morandini, J. S., & Smith, E. (2026). Demographic and clinical features of early-onset versus late-onset gender dysphoria among adolescent and young adult birth-assigned females. Archives of Sexual Behavior. https://doi.org/10.1007/s10508-026-03571-6
