Bulletin Number One 1983

all the three years of undergraduate studies. Our programme, which is geared to meet practical needs in the Hong Kong society, has placed slightly more emphasis o n application than theory. Courses like Data Analysis, Applied Regression Analysis and Applied Multivariate Analysis are ‘applied' rather than 'pure' Statistics. Since our Department is new, there is of course room for improvement as it develops. On balance, I think that the programme is just right for the present stage. Q. What is the relationship between Statistics and Mathematics? A. Statistics has long become an independent discipline and is not a branch of Mathematics. However, they are very closely related. The main difference between the two is in the way of thinking: Mathematics mainly involves logical thinking but Statistics involves both logical and non-logical thinking. For example, induction but not deduction is used to construct statistical models from available data, but drawing conclusions from these models still employs mathematical methods. Q. Could you explain to us laymen the way to construct a model? A. The statistician will have to do some educated guess work if he has to construct a model purely from the data and no additional information is available. Of course, it would be the height of our creativeness i f we could use our imagination to construct a model never conceived before. However, normally we would draw on past experience and use old models proven successful, such as the linear models, which are most frequently employed. So we can say that Statistics is at the same time a science and an art. Q. Could you enlighten us on the origin of Statistics? A. The mathematical basis of Statistics is the theory of probability, which is closely linked with games, gambling and divination (God). The frequent discovery by archaeologists of early forms of dice made from ox-bones by men's early ancestors testifies to the long history of the above activities. However, it was not until the Renaissance that knowledge about the operational laws of gambling began to be developed. And it was in the 17th century that these laws were formally stated for the first time in the precise language of mathematics by two famous French mathematicians, Pascal and Fermat, in their correspondence. The modern development o f probability theory started in the thirties of this century with the Russian mathematician, A.N. Kolmogorov, playing the leading role. (He is still alive today.) The subject is still undergoing very rapid development. As for Statistics, it may be said that it was first developed in Britain. For nearly a century previous to 1750, Britain had experienced relatively stable government, which had permitted the development of trade, industry and agriculture. She was thus much more rapidly and completely industrialized and modernized than any other country. It therefore seems natural that Britain should become the cradle of Statistics. The word 'Statistics' itself is the product of the late 18th century and the early 19th century. Some of the early applications of Statistics were (1) in actuarial science, e.g. the design of the life table by John Graunt in 1662; (2) in Epidemiology, e.g. the study of the spread of diseases; (3) in Economics, e.g. the study of business cycles; and (4) in agriculture, e.g. the effectiveness of fertilizers. Q. When was the development of Statistics accelerated? A. It was not until the early 20th century that Statistics was fully developed, again in Britain. Famous British statisticians at the time included K. Pearson and his son E. Pearson, both professors at the University College, University of London, Sir R.A. Fisher, Weldon, Galton, and Greenwood etc. Through the efforts of these pioneers ， great strides were made in the discipline. In particular, Sir R.A. Fisher, sometimes known as the Father of Modern Statistics, set up the theoretical basis for statistical inference. Q. At present, which countries are most advanced in Statistics? 18 RECENT DEVELOPMENTS