JA's Stereophile test lab seems like a well-equipped and PURPOSE-BUILT facility to conduct a test as I'll propose below.
Some have noted--in their portable audio devices that can play back multiple formats--that FLAC seems to drain battery run time faster than WAV. This may be due to added processing demands on processors having to decompress on-the-fly. (And these processors do get warm, like pwr-hungry std. PC and laptop CPUs. Some in-the-know may recall how warm those classic current-hogging Philips digital-audio ICs got, e.g. SAA7220, etc.).
Concomitantly, might these harder-working processors be spewing out noise -- via dirtying up the rail power -- and thus degrading SQ ... Unlike playing back uncompressed formats like WAV? Hence, some claim that WAV has an important edge in SQ.
This would mean that o'scope probes would have to be placed on a running device (and on the teeny, tiny Vcc pins of various ICs). Think Stereophile labs has been doing such tests for its Measurements section, so I assume Stereophile (JA's lab) has a relatively well-established (= repeatable, standardized) scientific procedure to conduct such a test.