I think the temperature change in the blade from a warm or hot water rinse is relatively small compared to the errors associated with the human element (stropping technique) and the sample population. Compound these issues with the tester bias, and lack of calibrated testing to ensure any meaningful and quantifiable results, and I think this notion may never be proven (unless for some reason a person wanted to test it in a lab...and getting funding for that might be difficult:D)
Simply put, my ability to strop consistently and gauge sharpness AND meaningfully compare it to a perceived sharpness hours prior is not nearly constrained enough to say anything on this subject that would amount to more than a hunch/suspicion/prejudice.
As always, YMMV;)