This paper is from Session 7B: Statistics and sports
which comes under Topic 7: Statistics education and the wider society
(Wednesday 14th, 11:00-13:00)
Teaching statistics to sports scientists and sports administrators: would it be easier to train monkeys?
Presenter
- Christopher Barnes (AIS Performance Analysis, Australia)
Abstract
Sports scientists are not generally renowned for their statistical acumen: even more so sports administrators. The problem is that some statistical concepts are essential to sports science, and intrinsic to some forms of sporting competition. Then there are the issues of fairness and transparency in athlete and team selection, performance analysis, talent identification (TID), and in determining whether particular training “enhancements” are in fact not harmful.This talk will present a number of case studies, based on elite sport, that illustrate the idiosyncrasies and difficulties of communicating valid statistical information to experts with limited statistical sophistication. The focus will be on techniques and tools that the author has found to be most effective in engaging and educating these experts, taking into account their perceived lack of time, brief attention span, suspicion and (sometimes) outright hostility, and their (unrealistic) expectations.
The resulting solutions are problem-specific, usually not intellectually satisfying and often pragmatic; but some broad directions and rules for successful knowledge transfer can be made.