There is some counter-evidence from m/s/In vivo studies of science labs and design teams
very similar to @dunbarHowScientistsThink1997's m/s/In vivo method