A Memorial for William N. Hurwitz

Abstract

T HIS memorial session of the Washington Statistical Society is to honor Hurwitz for his contributions to statistics. It is also an opportunity to honor him as a person and to acknowledge his contributions to the people who knew and worked with him. Those of us that had the privilege of working closely with not only owe a great debt for the opportunity of learning from him, and being inspired by him, but found in a friend who gave personal comfort and inspiration, advice and help, as few people are able to do. No matter how busy, or how much load he was carrying, he would always find it possible to spend time on someone's problem-statistical or personal-and he had great insight and ability in helping in both. Bill's approach to everyone was disarmingly informal and direct, yet warm and friendly. It mattered not whether you were the nation's highest authority in statistics, or a beginner, you were treated to the same warm informality-and to the children of his colleagues he was simply Uncle Bill and they loved him. All these personal characteristics, and more, won for the respect, confidence, and love that all of us felt so deeply for him. received his undergraduate degree in Mathematics, from Connecticut State College in 1932. In 1935 he received his Master's degree from Columbia University where he was one of the early group of students who became interested in and trained in statistics by Harold Hotelling. He was one of many who earned his way as he went through school during the great depression, and he never finished his Ph.D. His working career was primarily with the United States government, beginning in 1936 when he served as the statistician on the Attorney General's study of parole, working with Wayne Morse. He spent about two years on this study, and about two years with the Bureau of Home Economics of the Department of Agriculture. In September 1950 he joined the staff of the Bureau of the Census where he and I began a friendship and a team working relationship that was enormously stimulating to me personally, and quite productive. His contributions had many and important impacts on the development of sample survey methods, both theory and applications, and more generally on a wide range of developments in survey and census methodology and on the work of the Bureau of the Census. It also had impact on sample survey and census methods elsewhere throughout the world. Because my name comes first alphabetically, I appeared as the lead author in various publications of results, but our work was so much the product of both (and far more was produced than could have resulted from each of us working separately) that for most of it there could be no separation of credit. We stimulated each other enormously, with no real specialization of work except that he had a remarkable facility for solving problems in probability and mathematical expectation for complex