My understanding is that 1 and 2 are correct for constant SPL, assuming no cone breakup, and assuming we're talking about on-axis response instead of power response. The acceleration on the cone (change in velocity over time) goes in the other direction (increases with increasing frequency), if I recall correctly - which is why cone breakup sets in as we go up in frequency, rather than down.
I don't know what battery polarity test you're referring to in 3.