They showed the statistics based on their telemetry tools and said they match crash data.
Not that it was 100% from crashes.
Also the fact they can tell which one is in use does not mean that’s the reason it crashed. It could be crashes due to bad network handling or file corruption or something that has nothing to do with the GUI.
The ones that don't are more likely those who leave things on defaults, are involved with the project or a distro, or similar. No, I don't have anything that backs this up. The statistics they're using can never be accurate, by virtue of being free software that ships on privacy concious distros to privacy cincious people. There was a study that backs up this claim, but I'm not google.
OTOH, xfce is doing fine.
>For transparency, the one caveat in all of the above is that I've deliberately always focused on people using the latest Plasma release. We do still have a sizable chunk of users on X11 still using Plasma 5.27. Including them, the total Wayland adoption rate is about 76%.