I don't think you can really produce replicable scientific literature about this type of thing. Way too many variables.
We are already incapable of replicating studies with fewer variables.
We are already incapable of replicating studies with fewer variables.
A survey of company success (measured in profits, or market cap) vs office layout, trying to see if there is correlation would be more useful. Even accounting for the fact that there is a lot of cargo culting, and unhappy employees.