If 70% of chimney smoke re-entered nearby buildings, then those times when most buildings were burning, the air inside them would be nearly as smoky as what come out of the chimney. It's also very hard to believe outside of a tenement arrangement. In a village where there's lots of space between houses, and assuming all particulate matter fell back within the bounds of the village (itself a big stretch), it's hard to see how two-thirds of smoke would find its way back into buildings.
It doesn't mention over what sort of volume/area it re-enters... 70% of smoke from one chimney distributed over the whole neighborhood/town would be more subtle.
edit: "near-by" is a relative term
No, since one persons smoke could enter a dozen peoples houses.
"those times when most buildings were burning" would not have one burner per twelve recipients.
Fall back down to the ground and blow back into sealed houses, though? I think obviously wrong applies to the scenario he wants you to have in mind. Maybe the research has some specific conditions.
I'm not saying it's right, but it doesn't seem obviously wrong to me. Regardless, I'm sure whatever "research" there is must be based on something! A citation would have been appreciated so we could judge for ourselves...