That is, checking the correlation of people attending certain kind of
meetings -- this might help when trying to avoid allocating the same
slot for two equally interesting (to some significant group of people)
Haven't seen further comment on this suggestion, so I wanted to add my own 2c
that this seems like a useful idea.
Obviously, the chairs are asked what meetings not to be opposite, but they may
not know the quirks of their audience as well as a histogram from past
attendees might show. And yes, there'll always be some conflicts, but this
can reduce them.
Conversely, knowing the correlations might be useful if the secretariat ever
tried to proactively cluster related meetings closer together.