Many years ago, and I have no proof (I thought there was a write-up in AES or Audio but have not found it) so take it FWIWFM, I helped set up and participated in a test to determine the localization of subwoofers. It was a bear to set up, and we used a combination of in-room and outside (field out back) testing. Measurements were taken to level match subs to 0.1 dB (what a pain!) and keep the FR flat from roughly 40 to 120 Hz. The room was well-treated. We performed listening tests with a variety of people over several weeks. Among other things, we were trying to see how high/low people could tell the position of a sub. We used filtered pink noise and test tones.
The results were that virtually everybody could point to the sub at around 100 Hz, and essentially nobody at 50 Hz. In-between was a gradient, with 70-80 Hz actually about the point at which most people could no longer tell where the sub was in the room (or field). IIRC we went in 10 Hz steps, mainly to keep the test to a reasonable time. Similar tests were run in a few other places around the country.
I have run stereo, dual-mono, and mono (single) subs at various times in my system. I have always preferred stereo, but am sure at least part of that is because I was on the low end of the testing (around 60 Hz) and perhaps more because crossovers are not perfect "brick wall" filters. Crossed over at 60 Hz means there is still content at 70, 80, 100 Hz... When the test was run DSP was not around, and the highest-order analog filters generally in use were only 12 - 18 dB/octave.
Even dual-mono subs can be of benefit in a room since independent placement can help reduce the impact of room modes at the listening position.
FWIWFM - Don