“Oh Grandmother, what big teeth you have!” Incentives to spur scientific research at business schools have been treacherous

Abstract

Scientific research was introduced in business schools in the mid-1950s to raise the quality of business education as well as to improve actual business practices. Schools expected scientific research to provide a benign and respected path toward improvements, and government support for research and students’ strong interest in studying business brought financial resources and hiring opportunities that made large changes possible. Consequently, hundreds of business schools began urging professors to publish scientific research. Sixty years later, plentiful evidence is showing that scientific research has not brought expected benefits for business operations; on the contrary, business-school research has generated a high percentage of untrustworthy “findings”. This disappointing outcome has six interacting causes: (1) Business schools have expanded by hiring professors who study fundamental topics of economics, psychology, and sociology rather than topics relevant for business practices, so few research findings have been adopted and tested by firms. (2) Instead of evaluating the quality or consequences of research, schools have relied on evaluations by journal editors and reviewers, whose overriding criteria emphasize theory at the expense of practical implications. (3) Editorial evaluations have high error rates because editors and reviewers only know what researchers choose to tell them, and because the complexity of manuscripts causes editors and reviewers to disagree frequently. (4) Schools have incentivized professors by making promotions and job security depend on publishing numerous papers in “prestigious” journals. (5) Many professors have responded to these incentives by engaging in covert practices that make studies unlikely to be supported in replications, and some professors have gone so far as to manufacture fake data. (6) Researchers, editors, and reviewers have largely relied on null-hypothesis statistical significance tests, which are prone to interpret small variations or random errors in data as “significant” findings. Although there are fresh efforts to make research findings more reliable and impactful, the need for further interventions to reform the present knowledge-production system is all too apparent.