Aargh, headroom measurements... Those data do not support modders' argument. First, that difference is well within a driver's sample-to-sample variability as well as within measurement setting's variability (e.g., slight change in their position on the dummy head).
Second, if you change headphones' ear cup structure, of course, you will see some, most likely, small effect on their FR due to the change in the reflection profile. But this is very difficult to predict---a modder without measurement will simply make a haphazard decision. And more importantly, a slight change in the reflection pattern is unlikely to cause a difference adequate to remedy a large FR anomaly due to diaphragm resonances (like the Tascam's).