Old studies. studies that made assumptions about non responders. Studies where all the youth were nto transgender.
None of which change the fact that studies both with and without medical intervention the results are the same. Almost none of the kids who are trans desist. Deal with it.
Where are these updated studies that you're referencing?
From my understanding, both "sets of studies" that people on both sides bring out to support their respective positions have some methodology issues.
For instance:
The studies suggesting that "90% of kids desist" have the issue of using a very small sample size, and assuming that everyone who didn't come back to follow-up questionnaire was a "desist". Which artificially skew the numbers higher than what they actually would be.
Likewise, some of the more recent studies (like the ones commissioned by the Trans Youth Project) that suggest that "only 2.5% of kids desist" have some flaws as well. The most obvious one being that it's not an unbiased source (we wouldn't trust the Tobacco industry to commission a study to to evaluate the impacts of smoking).
Apart from that, the study was started over 10 years ago (at a time when far fewer kids were identifying as transgender - The UCLA study referenced
here outlines the sharp rise in youths identifying as such in the past 5-10 years), there's varying explanations for why there's been such a sharp recent uptick in the numbers. Some of it is likely due to society becoming more accepting, but one also can't deny that there's been a shift in the climate of certain social circles where it's considered "cool" or "trendy" to be anything other than cis/straight.
The study's cohort was also only including children parents who were willing to facilitate their transition.
The study also included some very young children, and only followed them for 5 years which presents some issues with regards to what we know about the learning patterns of very young children.
Those aspects will cause the numbers to appear lower than what they actually are for the desistence rate.
I would liken it to an example of trying to study "how many people can quit using Drug XYZ on their own, absent any outside intervention?"
If Study A used a really small sample size, and made the erroneous assumption that everyone who didn't show up for the follow up survey "must have been able to quit", then they would likely overestimate.
If Study B started before Drug XYZ became widely available, and only consisted of people who had family and friends who were willing to facilitate the usage, they would likely underestimate.
The studies indicating only 2.5% and all the way up to 90% are both likely missing the mark by quite a bit...but in the current social climate, it's going to be very difficult to get an objective study at this point...and even if it was possible, we'd have to wait for 5-10 years before the results came back.