Chance models: building blocks for sound statistical reasoning.

Abstract

A good understanding of chance models is crucial for mastering basic ideas in statistical inference. Mature students should be introduced to the concepts of inference through a study of the underlying chance mechanisms. They should learn to think globally, in models. In an introductory course, these models should have their own clear and unambiguous notation. Fuzziness and flaws, as encountered by our students in textbooks and software, may hamper their learning process seriously. The above claims are based on my experience as an instructor for university students. They should be substantiated by systematic research on the potential advantage of “thinking in models”, possibly also for younger pupils.