Nathan KEYFITZ, 1913-2010

Nathan Keyfitz had at least two major and distinguished careers, each of them having lasted at least 23 years, and with some “minor” careers wrapped into each. By most of us “lesser mortals” either of those two major careers would be considered as exceptionally successful. He is the father of survey sampling in Canada, and one of the founders and leading lights of mathematical demography in the world.

Born in Montreal in 1913, Nathan Keyfitz received his B.Sc. degree in the same city (from McGill) in 1934. In those depression years jobs were not easy to come by. At least local folklore has it that Dr. Keyfitz joined the Dominion Bureau of Statistics (the predecessor of Statistics Canada) as a coding clerk attached to the 1931 Census. At any rate, his first professional career started in the D.B.S. as a statistician in 1936 and more or less ended in 1959 when he left that organization.

During those 23 years, Dr. Keyfitz rose to the very top of the organization’s research activity and was appointed Senior Research Statistician — a post abolished after his departure. Professionally perhaps his most lasting contribution has been the establishment of the Canadian Labour Force Survey in 1947 — the first large-scale probability sample survey in Canada, and one of the first in the world. With incredible productivity, which characterized him in all his careers, he designed single-handed this survey, established the estimation methods, selected the sample, and provided the professional direction to this pioneering undertaking. The design he established right after the war was to last, with only minor improvements, for 20 years.

Still as part of his first career, he established sampling as a professional discipline at Statistics Canada. He hired several statisticians (not easy in those days!) and set out to make sampling experts of them — working as a one-man post-graduate department. He organized sampling courses to the non-experts, identified opportunities for sampling applications, and published several basic papers (e.g., on variance estimation based on complex samples — the first practical paper on the subject; and on updating sample designs without reselecting the entire sample). By the time he left the institution, sampling was firmly established there — in fact it became a world leader in the discipline.

During this first career he found time to obtain a Ph.D. in sociology (1952, Chicago), to serve as a lecturer in sociology at McGill (1948-51), spend three months as census advisor to the Burmese Statistical Office, serve as a member of the Indonesian Planning Bureau (1952-53) and pick up French, Spanish, German, Indonesian and Italian as his second to sixth languages.

His second and current professional career deals with population problems and was foreshadowed by his very first published paper (in 1937) in which he developed the 1931 life tables for Canada. In succession he was a Professor of Sociology at the Universities of Toronto and, subsequently, Chicago (1959-68), then Professor of Demography at the University of California, Berkeley (1968-72), and Andelot Professor of Sociology and Demography at Harvard. On a part-time basis he also accepted the post of Lazarus Professor of Social Demography at The Ohio State University.

In his prolific output during the last 25 years, he was one of a handful of people who set the groundwork for mathematical demography — but exploring along the way a spate of applied issues of great policy relevance to the world: family formation, causes and consequences of population change, urbanization, the middle class, pensions, the limits of population forecasting, the impact of census undercounting on the distribution of federal funds, etc.

In recognition of his numerous contributions, he was elected Fellow of the Royal Society of Canada, Fellow of the American Statistical Association, Fellow of the Royal Statistical Society, Member of the International Statistical Institute, Constituent Member of the Inter-American Statistical Institute, Member of the American Academy of Arts and Sciences and of the National Academy of Sciences. He served as Chairman of the Sociology and Anthropology Chapter of the Canadian Political Science Association, as Chairman of the Social Statistics Section of the American Statistical Association, as Vice-President of both the Canadian Political Science Association and the Population Association of America.

Above all else, Nathan Keyfitz is a scintillating human being with an enquiring mind.

Ivan P. Fellegi, 1983