Computationally intensive methods of statistical inference do not fit the current canon of pedagogy in statistics. Seven pedagogical principles are proposed to accommodate those methods and the logic underlying them. These include defining inferential statistics as techniques for reckoning with chance; distinguishing 3 types of research (sample surveys, experiments, and correlational studies); teaching random-sampling theory in the context of sample surveys, augmenting the conventional treatment with bootstrapping; and noting that random assignment fosters internal but not external validity. The additional principles are explaining the general logic for testing a null model; teaching randomization tests as well as t , F , and x-sup-2 ; and acknowledging the problems of applying inferential statistics in the absence of deliberately introduced randomness. (PsycLIT Database Copyright 1996 American Psychological Assn, all rights reserved)
The CAUSE Research Group is supported in part by a member initiative grant from the American Statistical Association’s Section on Statistics and Data Science Education