A critique of Tordoff et al. (2022)
Damn, THIS is what I'm paying for.
Thanks, Jesse, for doing the work that researchers and reviewers obviously aren't willing to do.
I bet you're often feeling like you're fighting against windmills, but you're doing stellar and really important work.
That's how it's done - not by mindlessly accusing everyone you disagree with of being a "groomer".
Man the failure to acknowledge drop-outs is SO terrible. You didn't even mention the biggest problem with it - why exactly would a not-depressed person who does not want blockers go to a gender clinic? This is BY FAR the biggest problem in my eyes. Like take the situation the authors seem to be describing, one where no amount of distress keeps you off blockers. In that situation you would logically expect anyone who wants blockers to end up on blockers, which just leaves the people who don't want blockers in the non-blocker group. If you are in the non-blocker group AND experiencing no distress the clinic is providing you literally zero services of value - it's not therapy because you don't need therapy it's not any medical intervention because no intervention is needed - so OF COURSE they dropped out!
Unfortunately, hard sciences once again have a bit of a Galileo problem - their conclusions are required to align with the prevailing views of society's designated Moral Betters (the Church in Galileo's day, and the woke recently), which means often the research methodology is reverse-engineered to generate the desired/obligatory conclusions.
As another example, the American Academy of Pediatrics (AAP) recently deleted sections from its website regarding early childhood development and facial cues for learning, which conveniently coincided with progressives pushing mask mandates for children.
Progressives love to claim that their political views are informed by The Science, but in reality it's the opposite: The Science is predetermined by mandatory adherence to progressive talking points. This study looks like the double-blind equivalent of the Texas Sharpshooter Fallacy.
A lot, like A LOT, of medical research is poorly done on the statistical side. The main critique I would have of this study is that the study design doesn’t seem to speak to the research question of interest (what is the effect of GAM on mental health?), and so they’ve tried to come around to it in a hacky way and not done a great job.
It’s common in studies of adolescent development to consider a flat trajectory a big success if the trajectory of a comparison group is going down, so I’m less concerned on that point. A lot of things get worse in adolescence and then recover.
I’m more concerned on selective attrition. It is absolutely not common to have 40% of your sample drop out within a year in a longitudinal study. It’s really hard to conclude anything with that kind of drop out.
The fact that you couldn't get these questions answered by the corresponding author, but she was happy to go on Science Friday to talk all about the study is frustrating. It was a given that she wasn't going to get any in depth questioning there though. Science Friday is a well-produced show but it never disputes research claims. Science outlets, even the good ones, are essentially science fandom at this point. They promote science and scientists. I work in science PR and most outlets run our press releases as is. Some will even slap a new byline on it without changing the content of the release at all. It's that bad. There is obviously great science writing out there (this piece shows it) but for anything that's not just waxing poetic about the wonders of science, or being critical of an easy target ( oil companies and the sort) you need to go other places.
Excellent article. As a layperson not well-versed in statistics, what stands out most to me is not even the difference in attrition rates between the two groups (which is a serious issue), but rather that they could have *6 people* left in one group at the end and still consider the study important. If you began with 6 people in the control group and had no dropouts, the study would be so small as to be practically useless. The fact that you began with 35 and ended with 6 makes it *worse* than useless, because of course the reasons for dropping out could be directly relevant to the question being investigated. Am I wrong?
Did anyone else hear a record scratch sound effect when they read “nonbinary youth who received gender affirming medical care"? This seems exponentially more troubling from a medical ethics standpoint than someone transitioning between sexes (which I think already has some medical ethics issues as Jesse has previously detailed), because being nonbinary (unlike transitioning) is literally nothing but -- and never can be anything more than -- a state of mind.
This is the gold standard for rigorous, non-ideological, science-based journalism. Why oh why can't we have more of this?
Obviously the non-GAM dropouts in this study all killed themselves, Jesse.
the traditional 'gatekeeping' of science was publishing in a recognized peer-reviewed journal. this meant the study was pre-evaluated by a board of scientists who have pertinent backgrounds, and then the scientific community responded to the study by way of letters, specifically critiquing aspects of the scientific method rather than not liking the results and/or the author. checks & balances. and yet we have an offshoot of the AMA doing politics instead of scientific inquiry. In the olden days, the Journal would lose credibility (heck, they wouldn't have published it in the first place knowing it would damage credibility). I no longer recognize the field of science.
Jesse that post is an act of courage as well as quite brilliant. Your analysis is incisive, it must’ve been tremendously hard work and you publish it at your own peril, you invite what will surely be a withering attack from the zealots.
Jesse’s type of inquiry keeps society out of trouble: deductive reasoning is our ONLY hope. Always and forever the bain of our existence, that is inductive reasoning; inductive reasoning is our most perfidy human intelligence and leads to that which we commemorate every November 11. Lest we forget - yes the sacrifice but more importantly the source, the stupidity of inductive reasoning, our great myopia, leading us inevitably and always to wanton destruction. The lesson of history: we don’t learn from history, we ALWAYS forget.
Bravo Jesse, you are a smart, courageous dude; as for your advisories, they bring darkness and violence. Standby.
This is the top shelf autism that I subscribe for. Well done. Going to have to sprinkle “variable dichotomization” into the mix next time I have to discuss data with someone.
Jesse, when you write "kids who SHOULD go on blockers and/or hormones", this implies that there is some definite methodology to determine which children should or should not receive GAM. What is this methodology?
It seems to me that such a methodology does not in fact exist. You own reviews of the available research (such as this one) demonstrate this.
Given that, outright banning of GAM for children seems to me to be wise public policy, and I support it.
And FWIW, I am a 66-year-old lifelong Democrat.
You're being extremely generous, Jesse. At every stage - evidence gathering, processing, reporting and socialisation, the paper's authors have at best failed to live up to standards, if not outright rigged the result.
And the lack of critical assessment by "science journalists" is approaching a fraud on the public - we all know a study with the opposite finding would have been picked over forensically, if not ignored. The "showtime" afforded to this is absolutely a function of its partisan usefulness.
This is exactly what I love from you Jesse, thanks for this great piece.
It would not shock you to learn that I think the methodology of science is it's most important aspect, and that papers like this (and so many others) that seem conclusion driven drives me wild.
I wish we could get to the stage where all studies of this sort could only be interpreted through pre-registration that includes the statistical models that are used to draw conclusions.
I need to bump up my contribution for this Substack. Thanks for putting in the hard work like this, Jesse. Someone has to do it.