Jerzy Neyman (1894-1981)

Born in Eastern Europe, Jerzy Neyman began his university studies at the University of Kharkov as World War I began. The teachers there had limited knowledge in current mathematics, so he augmented his learning by reading mathematics journals. Through his reading, he became familiar with the works by Henri Lebesgue whom he grew to admire. Later in life, he had the opportunity to meet Lebesgue who was gruff and walked away from their conversation while Neyman was still talking to him. Neyman was hurt, and he used the experience to ensure he would always help young students and encourage their enthusiasm.

Neyman became a lecturer at The University of Kharkov. Over the summer of 1921, he visited northern Poland as part of a political agreement at the end of the Russian-Polish War. He worked at the National Agricultural Institute and published papers about his work in agricultural experimentation, mainly regarding study design. These papers were part of the reason Neyman visited Karl Pearson at University College London (UCL) where he met Egon Pearon and began corresponding with him and later moved to London to join the faculty at UCL. In 1937, he visited and lectured at several universities in the United States. He was subsequently offered a position at UC Berkeley, and moved there to join the faculty and start the Berkeley Symposia on Statistics and Probability.

With Egon Pearson, he developed and published an alternative to Fisher's significance test. While Fisher's work uses only a null hypothesis and focuses on p-values, Neyman and Pearson included an alternative hypothesis and highlighted type I and type II error. Neyman also worked on creating an interval for parameter estimates. Probability is used in the construction of the interval, but the interpretation is ambiguous since the parameter is fixed. Neyman tried to avoid this confusion by calling the intervals confidence intervals rather than using the term probability.

Ronald A Fisher continuously denounced Neyman's work throughout his life, including where Neyman added to or discussed Fisher's own work. Once, Neyman presented his research in French with Fisher in the audience. When the presentation was opened to audience discussion, Fisher unexpectedly did not comment or challenge Neyman. Neyman later discovered that the reason for Fisher's unexpected reticence is that Fisher did not speak French. Despite what Fisher might say about Neyman's work, their methods for hypothesis testing have since been combined to form the hypothesis testing taught in introductory statistics. From Neyman and Pearson, there are typically two hypotheses, and from Fisher, a p-value is constructed to determine if the results are statistically significant.