The research is uncontroversial by now (many studies point to the same conclusion).
The trick is maximising those precious 40-50 hours and I think that's where we've gone up and down, backwards and forwards as an industry over the years. We've ended up with a lot of shoddy tooling and unnecessary complexity with our programming paradigms.
Only in the past year have I managed to figure out a way to work 50 hr weeks with the right tooling/programming language and have increased my output! And only by meeting someone else who had made the same mistakes I was making and then helped me discover what I was doing wrong.
Quite enlightening when I was on the verge of quitting the industry because I just kept thinking "this has to be so much simpler than how I'm doing it" and not really knowing how to make it so.