Yes, but are there other aspects that affect our perception of sound quality the audio analyzer doesn't measure? Or do some things like harmonics that are seen as distortion, actually improve your enjoyment of the music? If you have heard a good tube amp, that is a product that measures worse than a solid state amp but sounds better.
Or... the ears aren't nearly as good as some people make them out/think/believe/expect/are sure they are and perform well below the measured values.
With all tube amps (your example) the distortion becomes higher with increased output level but does not perform that bad at lower output levels. This (low output levels) is where they are usually used at the most. 1% of 2nd harmonic only at some peaks in the lows isn't audible that much.
In this case measurements do not tell the story as quoted distortion levels at a certain frequency range and specific output power say nothing about the distortion levels at various output powers unless plots are used with output power and frequency in one 3D plot (do they exist ?) or at various frequencies but spread over various output levels.
In any case one needs to interpret the values correctly and not just state ... this measures worse but sounds better as it is unclear how 'sounds better' has been concluded. I mean sounds better as in sighted test ? test over several days ? Truly blind test ?
Then there is a 'limited' bandwidth (which makes it measure poorer) of transformer coupled tube amps but these usually still measure better than many speakers/headphones do. Those owning speakers or headphones that roll-off well before the amp does won't notice any change in roll-off of amps anyway.
It's all about knowing your own boundaries, the borders at which you find out for yourself what is audible or not instead of accumulating (subjective) reviews and consensus.
Requires a LOT of time, some objectiveness and rigor in your tests and subjective listening combined preferably with known to be good performing gear.
Also one needs to draw the correct conclusions from those tests which is not possible without the aforementioned objectiveness and rigor.
Things aren't as simple as looking at some numbers and drawing 'a conclusion' that ears might be 'better'.
That conclusion may not be the same as someone else may draw.