I believe there are situations of placebo effect, but here some of Almarg's thoughts on the matter...
(quote used w/o permission)
Keep in mind that the swap will result in different frequency components flowing in the + leg and the - leg of each two-conductor pair of wires. It seems to me that would in effect amount to a change in the inductance of the cables, because the magnetic fields associated with current flow through adjacent conductors would no longer be equal and opposite.
As you probably realize, depending on cable length, cable type, and the impedance of the speaker at high frequencies, speaker cable inductance can sometimes have a significant effect on upper treble response. And upper treble response would seem to be something that can subjectively correlate with soundstaging.
The connection swap would also reduce the degree to which any noise pickup that may occur is common mode. I would not expect that to generally be significant in an amplifier-to-speaker connection, because of the low impedances and relatively high signal levels that are involved, but I would be hesitant to totally rule out the possibility.
As you know, I am certainly one who tends to be skeptical about a lot of tweaks that are reported, but in this case I would not rule out the possibility that the differences were real.
Best regards,
-- Al
(quote used w/o permission)