I've seen people do a study like that the right way. It works by having separate researchers invent categories independently and then compare them once they have classified everything. No talking about the project is allowed until everybody is done.