If someone gave me a few million dollars and a very cooperative ethics research board, these would be the studies I’d run:

Physical dysphoria study. I would like to put a bunch of cis men and cis women on cross-sex hormones for one or two months: a time period long enough that psychological effects would appear, but short enough that most of the changes are reversible. I’d be interested in seeing if a significant subpopulation of cis men and women felt like their bodies were “weird” and “wrong”, even without psychological changes; I’d also be looking for higher levels of dissociation and depression.

The biggest issue here is that I’d expect trans people and cis-by-default people to be much more likely to sign up for a cross-sex hormone study. So I’d probably describe it as “a test of the effects of hormones on well-being” without mentioning gender at all, although of course the subjects would be informed of the risks (hirsutism, libido changes, erectile dysfunction, a small chance of infertility).

Social dysphoria study. I have put considerable thought into how one would test for social dysphoria in a cisgender population. There are two problems: first, being read as the wrong gender is often considered insulting, which is a huge confounding variable; second, brief misgenderings are unlikely to upset even cis people with gender identities that much, because (in most cases) their genders are validated overall by society. But I think I’ve worked out a plan that should work.

There would be about three or four sessions, each five or six hours long, spread over the course of a few weeks. The subject would dress in cross-gender clothing, then spend the session engaged in ordinary social interaction with confederates. The confederates would be informed that the subject is a pre-everything transgender person; you’d probably have to get confederates that were very trans-positive and very good at using the correct pronouns. The subject would be informed that the study is of, say, people’s willingness to call a person a gender that isn’t the same as their assigned sex at birth. I would predict that subjects would find the experience very distressing and aversive; they would experience higher levels of social anxiety and discomfort.

Longitudinal study. Offer free hormones and surgery to a group of trans people, using an informed-consent model, on the condition that they follow up regularly.

There would be two primary goals of this study. First, there really isn’t enough study of long-term effects of hormones; a longitudinal study would let us figure out, say, if taking testosterone leads to an increased risk of uterine cancer. Second, a longitudinal study would allow us to figure out what characteristics are correlated with detransition. Particularly, I’m interested in what percentage of detransitions are a product of lack of social support and experience of transphobia, what percentage are a product of religious conversion or becoming a radical feminist and thus no longer believing that it’s okay to be trans, what percentage are people who are happy and satisfied with the effects of hormones but don’t want to be on them anymore, and what percentage are non-gender-dysphoric people who made a mistake. In addition, I want to figure out what traits allow us to predict who’s in the last group.

Puberty blocker RCT. Identify children who identify as a different gender, using the current criteria for gender dysphoria, which distinguish it more clearly from childhood gender-non-conformity. Assign half of them to take puberty blockers and half of them to go through the puberty of their assigned sex. Well-being is, of course, the most important measure here, but I’m also interested in what percentage of trans children who have taken puberty blockers become trans adults, and what percentage of trans children who haven’t taken puberty blockers become trans adults.

Autism and gender dysphoria. There are very suggestive studies about a correlation between autism and gender dysphoria; unfortunately, all of them involve tiny sample sizes. I want to enlist enough gender dysphoric people that we can actually figure out if those two things are correlated, or if it’s a fluke.