EPR Approach to Intro Stat:
Novelty, Practicality

In an e-mail message to me on May 12, 1996 (which he has kindly permitted me to quote), Bert Gunter (bgunter@njcc.com) wrote:

I agree with what you say [in two papers about the entity-property relationship approach], but do not understand what is novel.

The intuitive nature of the entity-property-relationship approach can draw attention away from its novelty and practicality. These can be seen, however, by considering a brief argument:

1. Most empirical research projects across all branches of science and throughout all branches of social and commercial endeavor can be usefully viewed as being studies of relationships between variables.
2. Almost all of the currently popular statistical procedures can be usefully viewed as being procedures to aid in the study of relationships between variables.
3. Therefore, it is practical to build the introductory statistics course around the concept of a relationship between variables.
4. The entity-property-relationship approach to the introductory statistics course is tightly built around the concepts of entities, properties of entities, and relationships between properties of entities (relationships between variables).
5. Other currently popular approaches to the introductory statistics course focus at most only peripherally on the concepts of entities, properties, and relationships.
6. Therefore, the entity-property-relationship approach is novel.

Most readers will agree that IF the antecedent propositions (1, 2, 4, and 5) of the above argument are true, then the conclusions (propositions 3 and 6) will also be true. But some readers may question the truth of the antecedent propositions, especially 1, 2, and 5. I invite these readers to propose counterexamples to this newsgroup that appear to refute those propositions.

I maintain that (while there are some questionable cases) if one considers a large number of individual cases (of empirical research projects, statistical procedures, and approaches to introductory statistics), one finds that propositions 1, 2, and 5 are well supported.

I believe that the main benefit of the entity-property-relationship approach is that it makes the vital role of the field of statistics in empirical research substantially easier for students to understand. The approach is described further in material at

http://www.matstat.com/teach/