The Training of Statisticians

Abstract

Since the fall term in 1946 members of the teaching faculties of the Department of Statistics in Chapel Hill and the Department of Experimental Statistics in Raleigh have been involved in a cooperative teaching program which has provided a unique opportunity to study the problem of teaching statistics to students who were candidates for a degree in statistics as distinguished from students who have an interest in the subject principally for its utility in certain special fields of knowledge. Due to the administrative organization of the Consolidated University of North Carolina, the Department of Statistics in Chapel Hill is mainly concerned with the theoretical aspects of statistics and has no direct connection with other departments in the University nor responsibility for teaching courses for any but its own students. In Raleigh the work of the Department of Experimental Statistics is less specific and includes teaching and consulting responsibilities to the rest of the College. Students who plan to specialize in theoretical statistics and minor in mathematics therefore register in the Department at Chapel Hill, while those who are principally interested in statistical methodology register in Raleigh and minor in a department in which statistics is applied. Students on each campus take courses on the other campus, i.e., students in Raleigh travel to Chapel Hill to take courses in Statistical Inference, Multivariate Analysis, etc., and students in Chapel Hill go to Raleigh for courses in Advanced Experimental Statistics, Design of Experiments and Sampling. However, evidence began to accumulate that students interested mainly in applications were beginning to exhibit lacunae in their knowledge of statistical theory while'a complementary condition was manifesting itself in the theoretically inclined students. This led some of us to reexamine our definitions of what a degree in statistics should mean, and in 1951 a Curriculum Committee was organized on an Institute of Statistics basis, and made a preliminary report to an Institute staff meeting in December 1951. The Committee agreed that the training of all candidates for degrees in statistics should be uniform up to a certain point. It was therefore their task to determine a minimum level of statistical theory and a minimum level of statistical analysis to be required of all degree candidates. It is perhaps worth noting that the name