Who should teach statistics?

 

“Most people are taught to drive by another driver, rather that by an engineer or mechanic, and it is perhaps reasonable to assume that a fellow statistics user is the best person to convey the principles. A certain amount of extra knowledge is useful, to explain the basis of concepts such as sums-of-squares; for example, but the basic procedures and, equally importantly, the “rules of the road”, can be taught by a non-specialist. Indeed, a mathematician may not have the necessary insight into a biologist’s mind to appreciate the degree of simplification required. A biologist may be better equiped to understand this, and also to know the diversity of backgrounds from which a biological course draws its students.”

 Kelly M. 1992. Teaching statistics to biologists. Journal of Biological Education 26: 200-204.

 

“Perhaps the greatest advances in statistics over the past century have come from those whose interests and training have come from outside the field, e.g., Galton, Gossett, Box, and Tukey.”

  Fienberg S.E. 1994. Invited comment. American Statistician 48: 71-72.  

 

The statistician and the scientist/technologist will find more in common, to the advantage, I believe, of both, if their education can be more closely linked.

Firstly, the statistician: during training (s)he should take part in a designed study or experiment, becoming involved in the design, the planning of data collection, the actual collection of the data themselves, and the subsequent analysis. This in turn will involve becoming au fait with some relevant existing theory in a particular field of application together with the previous data. [...]” 

Nelder J.A. 1986. Statistics, Science and Technology. Journal of the Royal Statistical Society A 149: 109-121.

"In conclusion, we again note that statisticians are naturally inclined to emphasize the formal aspects of inferential methodology, such as using efficient or unbiased estimators. The basic underlying assumptions, discussed in this article, are well known, but in the past have often been inadequately communicated. We urge a greater recognition of this highly important "soft side" of statistics in our training of students and in our support of clients." 

Hahn G.J. & Meeker W.Q. 1993. Assumptions for statistical inference. American Statistician 47: 1-11.

 

"La estadística es, con mucho, demasiado importante como para dejarla por completo a los estadísticos.”

George BOX p. xiii in Johnson D.E. 2000 Métodos multivariados aplicados al análisis de datos.

 

Quotations Relating to Tenure and Evaluation and to the Importance for Teaching of Real Data, Labs, and Consulting

 

Back - Home Page EGB