Too many variables, IMHO (dev experience, team size, technology stack, etc.). Startups in particular are inherently volatile places, and basing such a study on

Too many variables, IMHO (dev experience, team size, technology stack, etc.). Startups in particular are inherently volatile places, and basing such a study on them is fraught with problems. Most notably, the measure of "success" in startup-land is not irrevocably tied to code quality. Plenty of projects have succeeded despite their engineering and not because of it.