We agree.
Assuming it's speaker sample consistency, then it becomes pretty pointless to choose between models based on one measuring closer-to-ideal than the other, beyond a certain point. And, based on the M2 differences we are observing, that 'certain point' is pretty rough, really.
You're right, but I don't really think it's manufacturer consistency exactly, perhaps I should have rephrased. After all, we saw earlier in this thread both the left and right units measured identically in the bass.
Instead I think the most likely explanation is that for whatever reason the official spinorama made during the development phase is not 100 percent reflective of the products people are getting. I wouldn't be totally surprised if the spinorama was made during near-final development of the speaker, but some slight thing was tweaked near or after shipping. We also saw a similar thing happen with Buchardt when they slightly changed the crossover of the S400 and and the official measurements were not reflective of the changes (they've since updated those measurements).
So the mystery remains but I see no reason to believe anything is amiss with Erin's measurements. It is ultimately an EQ problem as you say, but why the EQ is wrong is the real question.
Without getting at least one or two more decent measurements of the M2, it'll be hard to tell. A single quasi anechoic measurement would be enough to tell. Heck, not even that -- I just need a nearfield woofer and port measurement to see if the same issues are present as with my earlier nearfield sum of the speaker being tested.