Yeah, this is why I said it devalued the usefulness of the stat.
It goes both ways though. There are people who checked that they use Ubuntu when they're on Mac or Windows the majority of the time, but there are also people that checked they're on Debian which isn't feeding into the Ubuntu top line number. I'm not sure which way this is going to bias it -- I could see the number being much higher or much lower. Debian was at 9.8% (iirc) for instance, and I don't imagine there are many people using both Ubuntu/Debian for desktop usage.
It goes both ways though. There are people who checked that they use Ubuntu when they're on Mac or Windows the majority of the time, but there are also people that checked they're on Debian which isn't feeding into the Ubuntu top line number. I'm not sure which way this is going to bias it -- I could see the number being much higher or much lower. Debian was at 9.8% (iirc) for instance, and I don't imagine there are many people using both Ubuntu/Debian for desktop usage.