Lots of good reasons for keeping 27" rims on bikes that came with them, but I'm skeptical that anyone can subjectively differentiate between otherwise equal wheels that differ only in diameter by 8mm out of 730mm (about 1%). To be convinced, I would need to see data showing the riders' ability to identify a 1% change in diameter as an isolated independent variable, in a statistically significant, blind experiment. I might even insist that moment of inertia be held constant, and would likewise concede that total bicycle mass be held constant also, with appropriate ballasting as needed to normalize these parameters.