The only thing they measured was that the thrust vector was decoupled from the cavity orientation and did not decrease when power to the cavity was heavily attenuated (fig 10). They therefore concluded that the forces measured are not thrust from the EmDrive. Everything else is (likely correct) speculation.