But if everyone is giving having the same effect on the fight then most of the feedback is likely misplaced, or misinterpretted because people focus on 1 thing and miss the bigger picture.
I agree with most of your post, but this bit stands out to me as something... I disagree with (for lack of something more eloquent).
The data they appear to be getting, assuming that they view all classes as more or less equal, does indeed seem to reward classes equally for their mere presence. It does not, however, take into any account the actual "effort" required, which I think is what many people misrepresent as class imbalance.
Can anyone honestly say that playing a Pyro Powertech or a Rage Juggernaught is in any way as challenging as playing a DPS Merc or DPS Sorc? I haven't played my Merc in PvP, but I've done the other Three (or their mirrors) and don't find it in any way comparable.
My personal belief is that people who are just not as good tend to look up which classes are "easier" to do well with before even rolling, or if they start with one of the harder ones, they are much more liable to give up part way through and reroll. This leaves a higher proportion of those complex AC's being placed by more skilled players, and a larger proportion of less skilled players playing the "easier" AC's. This affects all the averages, skewing any metrics gathered and hiding underlying class inadequacies or excesses.
I do not think things are totally balanced as is, but I, as with almost all the community, do not play all 8 AC at the same gear and skill level, and so am a biased judge. As such, no changes that affect "total output" should be based solely on community feedback. That being said, community feedback (and views on classes that are under or over powered) is usually quite keenly correct in terms of what classes could do with balancing changes that either simplify an AC's rotation/gameplay mechanics or add more complexity to them.
This is something I feel developers often don't pay enough attention to.