Designing Experiments That Generalize

Organizational research has relied too heavily on methods characterized by passive observation, likely because there is a widespread belief that experimental research has limited generalizability. However, this is often because researchers (and reviewers or editors) misunderstand the nature of generalizability and what it requires. This article reiterates the importance of experimental research for understanding organizational phenomena and separates the legitimate concerns about experimental generalizability from the irrelevant ones. Whereas most criticisms of experiments focus on sample characteristics and mundane realism (i.e., superficial resemblance to the real world), more attention needs to be paid to the degree to which the treatment manipulation is valid, representative, and strong.