Randomizing lateral speaker positions averages out width-mode influences in tests

Evidence

In these experiments, loudspeaker position effects required multiple listening sessions with each loudspeaker auditioned in each position, and the results averaged. In each of the rooms, the room modes associated with the length dimension would remain relatively constant during the comparisons, with the width and tangential modes varying as the loudspeakers were randomly moved among the three (lateral) positions during the test. Residual differences would be averaged out to some extent, so even the low-frequency modes, which are a significant factor in normal listening, would be attenuated in their influence in these results. — Toole, Floyd E, Sound reproduction- the acoustics and…, 554