Thread by @LeorSapir

IMPORTANT UPDATE

4 days after my fact-check of @jack_turban's article reviewing studies in @PsychToday came out, Psych Today issued a correction to Turban's article (or asked Turban to do so).

🧵

The new version is here: psychologytoday.com/us/blog/politi…

4 corrections were issued, along with 2 stealth edits:

* Kaltiala et al 2020: hormones are "not enough to improve functioning and relieve psychiatric comorbidities among adolescents with gender dysphoria." BTW--this sentence is in the abstract, not buried in the article.

As I mentioned in my fact-check, Kaltiala herself emphasized to me that their study should not be used to claim that hormones reduce suicidality. That is a stronger caveat than the one Turban now acknowledges.

* van der Miesen et al 2020: the authors note they found no "direct evidence" of mental health benefits from puberty blockers. The correction is accurate enough.

* Achille et al 2020: Turban acknowledges that the improvement was only in MtF and only for puberty blockers. As I pointed out, however, this group was 1/3 of the total cohort studied, and represents a minority of those referred for treatment these days. Why not mention this?

* Turban et al 2020: Turban now acknowledges that his own study did not control for psychotherapy as a confounding variable, meaning it's possible those who reported better mental health did so for reasons having little to do with getting hormones. This is accurate enough.

Stealth edit #1: Allen et al 2019: Turban now writes: "Of note, the adolescents also received psychotherapy."

Stealth edit #2: as I noted in my fact-check, Turban dismisses studies that found no improvement in mental health when they have low sample sizes, but includes studies with small (or even smaller!) sample sizes when they support his conclusion.

Turban and Psych Today stealth edited out of the original article the sentence "likely due to their small sample size" from the discussion of Carmichael et al 2021.

It's unclear why Psych Today didn't ask for corrections to Turban's interpretations of the other studies--or, if it did, why Turban dismissed the request. For example, the Dutch study's flaws are by now well known, and by any reasonable measure seriously weaken its conclusion.

Likewise, as the authors of that study themselves recognize, it is very likely not applicable to the majority of cases referred for treatment at gender clinics these days. Given the stature of the Dutch study, surely this is relevant background Turban could have provided.

Turban didn't correct the misrepresentation of the Green et al 2021 study either. The authors explicitly note that causal claims about hormones couldn't be inferred as psychotherapy may have been a confounding variable. Why acknowledge this in his own study but not Green's?

Also, Turban's discussion of Tordoff 2022 continues to mislead. He fails to note that there was no stat significant improvement in the hormone cohort and even among the non-hormone cohort (whose deterioration led some to believe hormones work) 86 of the 92 were lost to follow up.

So let's take a final tally here. Of the 16 studies:

3 studies are inapplicable to most teen referrals today
4 showed no improvement
5 had confounding variables
3 showed mixed results at best
1 showed positive outcomes but relied on heavily biased sample

Which brings me to the biggest correction that should have, but didn't, happen: Turban continues to say that 16 studies show, on the whole, that hormones "result in" better mental health. That's very misleading causal language. At most, he should have said: "coincide with."

For my full fact-check of Turban: realityslaststand.com/p/the-distorti…