They likely outperform for the same reason they are self-thought— i.e. it's most likely a confounder. If you have enough curiosity/pull, motivation, discipline to do it, you'll outperform in general, self-thought or not.
There's also a survivor bias at play, in the sense that the self-teaching attempts results that you see in professional environments are those that made it.
So, the true question is, for those that tried this path and failed, would have they been better off with proper teachers. Sometimes it may just be a matter of time - taking longer to learn because of various reasons, that wouldn't be the case in normal educational environment.
With this said, I went through CS and math school and enjoyed it, but I don’t think I’m compatible with the format and prefer self learning.
Seeing as no statistics were even presented, I don't think you need to worry about statistical fallacies. Often is just used as a weasel word in this case.
There's also a survivor bias at play, in the sense that the self-teaching attempts results that you see in professional environments are those that made it. So, the true question is, for those that tried this path and failed, would have they been better off with proper teachers. Sometimes it may just be a matter of time - taking longer to learn because of various reasons, that wouldn't be the case in normal educational environment.
With this said, I went through CS and math school and enjoyed it, but I don’t think I’m compatible with the format and prefer self learning.