It isn't the dipole design, its the lobing that occurs on the vertical plane. All of that projected energy is interacting with itself, and it will produce a very complex comb-filtering pattern. The thing is, comb-filtering can be difficult to hear, so while the response might look like a mess, it may not sound like one. Indeed, I have heard maggies and other planars and electrostats that have this quality yet they always sounded fine.
Neither free-air or Klippel measuring systems would be able to disentable this problem. It's possible a far-field, multi-mic measurement might help though.