How would you study it exactly? Start-ups fail all the time for reasons that have nothing to do with their organisation or even execution. And if you prove that some succeed with it and some fail with it, as you suggest, then you've proven nothing.
My hypothesis is that it doesn't make any significant difference once the startup has grown beyond its founders (and they can work however works for them). No idea how to test that.
