Those statistics gained were studies done from four clinics before 2013, they were diagnosing GIDC (Gender Identity Disorder in Children) under DSM3, and DSM4, and seeing if they ‘grew out of it’ some years down the line. What makes it so inaccurate is using DSM3, and DSM4, any child that showed ANY kind of gender non conformity was diagnosed with GIDC, the children never said they were ‘trans’ in almost the entire studies. (DSM4 didn’t even need being uncomfortable with your body as a diagnosis). Two of the clinics could only diagnose children that were brought in by concerned parents. So if dad or mum didn’t like their son playing with barbie dolls and was concerned, they'd bring their kid to the clinic and would subsequently be diagnosed with GIDC under DSM4. Parents who let their child socially transition and were comfortable affirming their identity most likely didn't participate due to the nature of the studies. Since 2013, GIDC was renamed to Gender Dysphoria in DSM5, with a whole new diagnosis than GIDC under DSM3/4. As far as I'm aware, there are no studies that show desistance rates since. Edit: Didn't get to add pdf link of the commentary I read about this, and where I summarised from. Link was on old phone, and I was too lazy to find it. Here it is now: https://www.researchgate.net/publication/324808865_A_Critical_Commentary_on_Follow-Up_Studies_and_Desistance_Theories_about_Transgender_and_Gender-Nonconforming_Children It's a lot more in depth, further explains how those numbers are kinda shithouse.