William Gemmell Cochran lived from 1911-1980. He was born in the Royal Burgh of Rutherglen, Scotland. The Cochran family was a middle class family and provided William with a very normal childhood. His father was a railw ay employee all his life and William had one brother. Cochran was always good at academic subjects but not at art. He always had an intense desire to be first in his class. His repeated success in school won him many prizes and scholarships. He receiv ed scholarships at Glasgow and later at Cambridge University. This was an early start to a long successful academic career.

Cochran studied mathematics, applied mathematics, and statistics with Wishart, while attending Cambridge University. He was persuaded to go to Rothamsted without a Ph.D. and begin his practical work. During his stint at Rothamsted Cochran attende d lectures given by Fisher at University College. Cochran performed experiments, attending to weather issues, differential fertility of plots, and lack of replication. In six years at Rothamsted, Cochran completed 18 papers. Yates and Cochran worked on sample surveys while at Rothamsted, including a census of woodlands in 1938.

In 1939 Cochran visited Ames, Iowa and became a professor at Iowa State University. Raymond Jessen led sampling developments at Iowa State and was very grateful to have Cochran’s research strength there. During his time at Iowa State he served on and chaired an advisory panel to the U.S. Census. Throughout World War Two, 1943-1944, he worked with Samuel Wilks at Princeton. He worked in accord with the Statistical Research Group. Shortly after World War Two Cochran was persuaded to head the gra duate program in experimental statistics at North Carolina State College. In 1949 Cochran chaired the Department of Biostatistics at Johns Hopkins University. He wrote two books, Sampling Techniques and Experimental Designs, during his tim e at Johns Hopkins. In 1957 Harvard University established a Department of Statistics in which Cochran was hired on to add strength and senior leadership.

Cochran specialized in fields such as sample surveys, experimental design, and analytic techniques. The U.S. Public Health Service researched the effects smoking had on lung cancer and Cochran was the statistical representative. In 1963 the Advis ory Committee told the Surgeon General smoking was a direct cause of lung cancer. Cochran was responsible for discovering many statistical methods. Methods of including or excluding an independent variable in multiple linear regression were developed by Cochran in 1938 (Regression Analysis, 34). He also invented the Cochran Q-test, which is used to evaluate two variables that are measured on a nominal scale. Cochran performed many agricultural studies such as the Influence of Rainfall on the Yield of Cereals and the Field Counts of Diseased Plants.

Cochran’s knowledge in the field of statistics did not occur by chance. He studied and worked at the most prestigious universities. He was a colleague of and a student of some very well known statisticians. Cochran was a student of Fisher’s and a colleague of both Yates and Wilks. The statistical knowledge of Cochran was used by the Census Bureau, it was used in the medical field by the Surgeon General, and it was used in agriculture to determine crop yields. "Few, if any, have contribute d more to statistics as consultant, teacher, researcher and author" (Hansen).

Bibliography

 

Cochran, William G. Sampling Techniques. John Wiley and Son, Inc. 1997.

Cochran, William G. Contributions to Statistics. John Wiley and Sons, Inc. 1982.

Hanson, Morris H. Statistical Science, "Some History and Reminiscences on Survey Sampling." 1987.

Williams, E. J. Regression Analysis. John Wiley and Sons, Inc. 1959.