On the Misinterpretation of Insignificant Estimates

This article aims to remedy a common misinterpretation of insignificant coefficient estimates and presents a method to find hidden information in insignificant effects. I first discuss possible reasons for insignificant effects. I then use two distinct research areas as case studies: (1) how divorces are associated with changes in children's behavior to demonstrate how “counteracting effects” could cause insignificant estimates and to demonstrate a method to simultaneously identify positive and negative effects; (2) the “hot hand” in basketball to demonstrate how improper interpretations of insignificant estimates spawned decades of over-reaching conclusions in further research, best-selling books, and the popular press