My teaching has been influenced in a major way by the ideas set forth in the book “How Learning Works: Seven Research-Based Principles for Smart Teaching”. At the time of its publication, the book’s authors were all members of the CMU Eberly Center for Teaching Excellence, which still utilizes much of the material from the book. Two particularly useful links lay out and describe seven teaching principles and seven learning principles.

I think it is particularly useful in the training of statisticians to engender metacognitive skills. This Wikipedia page on metacognition lays out three types of metacognitive knowledge, and the third one refers to knowing when and why you should use certain strategies. I find that, in training practicing statisticians, this is essential – when approached with a specific dataset and associated question of interest, the statistician must assess the situation and make decisions about the analytical approach. When a student is taking a class on, say, regression models, he or she knows that the solution to any homework problem will be some type of regression model and they just need to decide which one is the most correct. When analyzing data from a client outside the confines of a specific course, a student must first decide which of several possible broader approaches are even appropriate, and the set of possible solutions becomes much larger. If we can help students think about their approaches to classifying problems and developing analytical strategies through metacognition, we’re training better statisticians.