This paper represents an extension to the search for methods more appropriate to exploration and testing of sociological propositions about organizations. This search has focused recently on the use of simulated organizations in laboratory experiments. Based on their experiences with a unique laboratory experiment on organization, the authors suggest that the advantages which normally accrue to experimentation cannot be generalized easily to experimentation with the miniature replica model of organizations. Following a report of the structure of their laboratory experiment, the authors attend to four critical problems associated with this form of experimentation: (1) the nature and combination of properties considered to constitute the miniature replica of an organization, (2) instituting system and measurement controls, (3) the provision of structural proof, and (4) randomization.