The compact guide to classical inference
Today we’re releasing the first of about a dozen installments of a new short book that emerged from the StatPREP summer workshops. The book, Compact Guide to Classical Inference, is available in an online format here. It’s intended to help instructors see that there can be a simple way to teach basic statistical inference.
In addition to the book, we’ve set up an online discussion forum through the Mathematics Association of America. You’ll need an MAA Connect account, but you can set this up through the link even if you don’t belong to the MAA.
The first installment includes the preface and the first chapter, “What is classical inference?” The standard answer is that classical inference is the set of formula- and probability-table based tests found in conventional textbooks, as opposed to “computer age inference” which uses computing, randomization, and simulation.
At the StatPREP summer workshops, I led breakout sessions called “It’s all regression.” In those sessions, I showed how the usual inferential settings of introductory stats–difference between two means, difference in proportions, slope of a regression line, ANOVA–are all really the same thing. Rather than needing four different chapters with different formulas, the same simple formula carries out the hypothesis test in all the settings. Similarly, confidence intervals are constructed with a single simple formula.
The “Compact Guide” lays out systematically how to do this.
We’re releasing it in installments in order to make it a quick read and in order to support questions and discussion, book-club style.