My data points mostly aren't from people I've hired, but from many people I've worked with or known from techie communities. So there's some survivor bias there.

Like much of the my engineering culture philosophy, it's based on things I've seen work and not work. There's no grand unified theory of everything, and we usually don't know exactly to which factors to attribute success or failure.

I also try not to be simple-minded judgemental, especially when it comes to excluding people. For example, someone experienced might have less conspicuous excitement, but still pride themselves on rock-solid effective execution, and we can also see that working out very well. Or someone might be coming from a ruthless stack-ranking deathmatch company, but that doesn't mean they can't think like a highly-effective collective if they're shown what that's like.