Improving the Climate for Statistics

Abstract

INTRODUCTION I am very honoured to have been invited to deliver the 10th Vidya Jyothi Professor V. K. Samaranayake memorial oration. I thank the Director and the Academic staff of the U.C.S.C., and the University of Colombo for this. I believe I am the first non-Sri Lankan to have been invited in this way. I understand this is not only the 10th anniversary of this memorial series, but is also being celebrated as the 50th anniversary of computing at the University of Colombo. Professor Samaranayake (Prof, from now on in this talk), influenced developments in Sri Lanka in many areas. I am a statistician and therefore I use this occasion mainly to discuss his influence in this area. In particular I consider some topics that combine statistics and computing. This influence between our organisations was partly via a link between the University of Colombo and the University of Reading. The formal link was more than 30 years ago, but some of the activities remain important today. In 1967,50 years ago, Prof first taught FORTRAN programming to University of Colombo students. This was the same year that I joined the Department of Applied Statistics at the University of Reading. Statistics has been a problem subject to many people, in many countries, for a long time. I describe this problem in Box 1 below. Box 1: Common problems with statistics teaching Service course training is often dominated by analysis, with relatively little on data organisation, or on design. Examples are usually relatively small, hence little time is devoted to important descriptive methods. Sometimes a "recipe-book" approach is used which usually results in topics covered in their order of mathematical complexity, rather than their importance. It also often results in the overuse, and often irrelevant use, of significance tests. This approach provides little understanding of principles