Teaching graduate (and undergraduate) econometrics: Some sensible shifts to improve efficiency, effectiveness, and usefulness

Building on arguments by Joshua Angrist and Jörn-Steffen Pischke arguments for how the teaching of undergraduate econometrics could become more effective, I propose a redesign of graduate econometrics that would better serve most students and help make the field of economics more relevant. The primary basis for the redesign is that the conventional methods do not adequately prepare students to recognize biases and to properly interpret significance, insignificance, and p-values; and there is an ethical problem in searching for significance and other matters. Based on these premises, I recommend that some of Angrist and Pischke's recommendations be adopted for graduate econometrics. In addition, I recommend further shifts in emphasis, new pedagogy, and adding important components (e.g., on interpretations and simple ethical lessons) that are largely ignored in current textbooks. An obvious implication of these recommended changes is a confirmation of most of Angrist and Pischke's recommendations for undergraduate econometrics, as well as further reductions in complexity.