We have two different theories here. We're smart people, right? We should be able to figure out if we just percieve software quality to be worse or if it really is.

So how do we measure this in some valid manner?