No, include autistic kids. Exclude unwanted kids regardless of anything, they'll skew the results. The unwanted neglected kids growing up with poppy playtime and skibidi toilet are going to program games for a job after being told to go away by their entire family. Everyone else wouldn't have used computers as often because they were spending time with friends and family.
My parents loved and cherished me growing up, and still do, which is part of why I was the only kid in sixth grade with a laptop (the other part was I had a disability accommodation with the school that allowed me to type my assignments rather than write them by hand). The fact that they encouraged my programming talent at that age, didn't get mad when I installed a Fedora dual-boot on that laptop, and bought me the book Python for Kids for my 12th birthday, is why I'm a programmer now.
I'm sorry your parents didn't show you the love and support you deserved, but that's not the criterion we should be looking for.
Actually that's also an interesting statistic to cover. What's the proportion of programmers who learnt because they were supported vs unsupported (and while we're at it do code quality analysis just to see)