You have really got to be more attentive than this when you're reporting things. This conclusion is not supported by the data.

Your attempt to reinterpret the data also has serious flaws, primarily in that you're breaking the longitudinal structure of the data. But the authors already gave us a better look at the longitudinal structure.

Can you please explain Figure 3 to us?

* What are the three categories--what do they mean, where do they come from?

* How many people are in each category, roughly?

* What would a chart look like that if you had a population that had a "strong trans identity that didn't change that much"? What do the charts actually look like?

* Within a factor of 2 or so, what fraction of the "often" group in the first study window ended up as "often" in the last window? What about the second (12-14) window?

* What is the significance, and how many people does this represent, of the dark segment of the chart disappearing in the middle panel?