I'm with DiegoFrogs on the controlled experiments question. 1.29% is a lot to ask from measurement capability in such a subtle and complex parameter as "smoothness", and not very easy to meet in a system with so many confounding and/or uncontrolled variables, e.g.: tire pressure, individual rim variations, individual tire variations, spoke tension, spoke length; to name the ones that first come to mind.
Speaking for myself, I wouldn't bother going after or worrying about a 1.29% difference in "smoothness" between 27" and 700C wheels. Now, if there is a difference greater than that being perceived or measured, it must be due to something in addition to or besides the difference in rim diameter.