News

Having just refreshed and expanded my Applied Statistics at my local University, I have two comments:

1. Really learning Statistics is hard work and takes time and commitment, so it is a multi-year process. In college and grad school there are competing commitments, courses in JAVA, Operating Systems, Compiler Design, Algorithmic Complexity etc., etc., etc. so unless somebody wants to understand Statistics in depth it will not happen. Further, if it is not done before starting work, it will be hard to do on the job when other job and personal commitments require time after the 8 hours/day are done. Therefore, in the workforce one might expect a shallower understanding of Statistics to be pervasive.

2. Unmentioned in Arnold's letter is the issue of the Probability Nazi, the "statistics" version of what Joan Baez called a "folksong Nazi". Although the mathematics of Probability and Statistics are politically neutral, there is a fierce political debate raging about the interpretation of Probability to the real world. There are six major interpretations, all of which have problems. Data miners tend to be more practical, so I imagine that if they come into contact with, say, a "Bayesian causal reasoning" zealot, they will confuse form and substance and ignore that good part of the discipline.

In general, however, I urge everyone to go out and learn more Statistics!

Dr. Lucian Russell Chief, Data and Knowledge Management Practice DynCorp Information Systems 1331 Pennsylvania Avenue NW Suite 730 North Washington DC 20004