Chris Carollo

@flounder2200 @Youknowthatin @mattyglesias Sure? I'm just describing what's making me skeptical of the study and in particular the proxy that they're using.

@CrazyTalk @Youknowthatin @mattyglesias Pretty sure governments, insurance companies, and other large stakeholders in the process are not going to be making decisions based on your gut.

Hardly a new story at this point but one that’s still making its way through American politics…

@flounder2200 @Youknowthatin @mattyglesias And, admittedly anecdotally, but "1-in-10 GD folks have severe psychological issues pre-treatment, but 6-in-10 do afterwards" just does not align with anything that I've actually seen.

@flounder2200 @Youknowthatin @mattyglesias My beef is mostly that people are reading this as proof that GD treatment causes harm, and while it may point in that direction, without knowing what the post-GD-presentation treatment is for, and what their outcomes are, it just seems like a inconclusive proxy to me.

@flounder2200 @Youknowthatin @mattyglesias I never said it was wrong, but it does raise a bunch of questions for me: - Why measure treatment not outcomes? - Why does the post-presentation, GR+ group rise to match the treatment rate of the GR- group? - Why is the pre-presentation treatment rate of the GR+ so much lower?

@CrazyTalk @Youknowthatin @mattyglesias "Why am I so eager to come up with a list of 'problems' with a study I haven't read, whose design I don't know?" To be blunt, this reminds me *exactly* of discussions of climate-science papers with self-described "skeptics". They don't know the science but they know it's wrong.

@Sebastian_Hols @flounder2200 @Youknowthatin @mattyglesias I agree there, mostly I feel like this study raises more questions for me than it answers.

@CrazyTalk @flounder2200 @Youknowthatin @mattyglesias A bunch of the US studies have 30-50% lack of followup after only three or four years. That introduces a bunch of confounding factors! The whole discipline is very sketchy.

@Sebastian_Hols @flounder2200 @Youknowthatin @mattyglesias It being the whole population of the country is great, but that doesn't prevent confounding systemic factors (which large sample sizes don't help with).

@CrazyTalk @flounder2200 @Youknowthatin @mattyglesias It’s meaningful because the system is literally the whole population of the country! If that’s going to be your problem you aren’t going to like any study ever from the US.

@Sebastian_Hols @flounder2200 @Youknowthatin @mattyglesias I guess I don't see why sample size is meaningful if it's a systemic thing (ie. if once people visit a clinic they have a more thorough examination, are "in the system" so more likely to have followup treatment, etc)

@CrazyTalk @flounder2200 @Youknowthatin @mattyglesias That’s the kind of objection that I totally believe could cause small effects, but it seems very unlikely, that all across the country, across a sample size that large, that is happening at the magnitude we are talking about.

@Sebastian_Hols @flounder2200 @Youknowthatin @mattyglesias Also, the thing I find most surprising about the data is that the "pre-clinic-presentation-but-do-undergo-treatment" quadrant has such low levels of serious treatment; all the others, treated or not, are in the 50-60% range. I'd love more investigation into that.

@Sebastian_Hols @flounder2200 @Youknowthatin @mattyglesias I didn't assume higher levels of scrutiny, I mentioned it as a possible confounding factor. Mostly I just dislike using treatment rates as a proxy: what were they treated for? what were their outcomes? etc.

@flounder2200 @Youknowthatin @mattyglesias Right, that's why I mean the "before state" -- was the same level of scrutiny and monitoring provided/given to the non-control cohort before and after first presenting to clinics? If not, the fact that they got more treatment after isn't really indicative of a problem.

@CrazyTalk @Youknowthatin @mattyglesias There was no selected “before cohort” of patients because all patients in the entire country were included both before and after. You don’t seem to have a very clear understanding of the study design.

@benryanwriter @mattyglesias I don't think that's true; she's saying that interventions don't change treatment rates. That's not the same as outcomes being identical. Which is kind of my whole complaint about using treatment rates as a proxy for things!

@CrazyTalk @mattyglesias Here she inadvertently argues that gender-transition interventions have no impact on mental heath outcomes.

@benryanwriter @mattyglesias Cool, I think her concerns would be good to actually address.

@CrazyTalk @mattyglesias She issued a 17-part skeet storm.

@benryanwriter @mattyglesias That doesn't really address her concerns either.

@CrazyTalk @mattyglesias The p-hacking claim on her part was just spaghetti thrown at the wall.

@benryanwriter @mattyglesias That's fine, but it, for example, doesn't really address the points that Alejandra Caraballo brings up.

@CrazyTalk @mattyglesias In Finland, specialist psychiatric care is typically reserved for those with more severe mental health problems.

@flounder2200 @Youknowthatin @mattyglesias Yeah I mean you're quoting the study I'm criticizing. The question is whether the "before" state is a reasonable control -- ie. do care levels change after treatment, are the "before" cohort randomly selected, etc.

@Youknowthatin @CrazyTalk @mattyglesias In Finland, it absolutely is a good indicator of severe psychiatric morbidity:

@arindube This chart only goes to 2019, but this was also true in the ~2022 tight economy too, right?

A striking fact I learnt researching my book: For those in the bottom half, much of the real wage growth between 1979 and 2019 happened over just 7 "tight" years. For top wages, full employment wasn't a big deal. For bottom wages, it was the nearly the whole game.

@noamscheiber This analysis seems to indicate that the *median* wage premium, while lower than the *average* wage premium, remained basically steady for the past 35 years (and is higher than the era before it): libertystreeteconomics.newyorkfed.org/2025/04/is-col…

A lot of people ended up with much smaller wage premiums, while a smaller number of people ended up getting much bigger wage premiums. The distribution looked more like so 4/

Wanted to take more time responding to the objection to my essay about struggling college grads from @mattyglesias, who notes the avg college wage premium is still very high today. That's true, but I think misses the big shift over the past generation. 1/ x.com/mattyglesias/s…

@K9Fiferr @besttrousers @JadeAtrophis You claimed "most [Somali immigrant] adults are unemployed" and that's not true. But clearly you just finding justifications for being racist and moving goalposts so I'm just going to block and move on with my day.

@CrazyTalk @besttrousers @JadeAtrophis Why would it be different in the United States

Funny how this only ever applies one way, eh