Monday 28 March 2011

Happy Birthday Karl Pearson

Sunday was Karl Pearson's birthdate, and I let it slip being busy with other things... But David Bee posted a reminder, as he often does with statistical birthdays, to the AP Stats EDG.  With grateful appreciation, I will steal his entire post:  (the picture above is the same, I believe, as the cover of Ted Porter's biography of Pearson.  )
---------


Karl Pearson, FRS, 27 March 1857 - 27 April 1935

Karl Pearson, who according to author and statistician David Salsburg, started the/a “statistical revolution” in the 1890s, founding the world’s first academic statistics department in 1911 at University College London, was born in London. Brought up in an upper-middle-class family, he was home-schooled until age 9, then sent to University College London, studying there until he was 16, at which time he left because of illness.

He was home-taught by a private tutor for two years, leading him to Take the Cambridge Scholarship Examinations in 1875. He came in Second on the Exams, winning a scholarship to King’s College.
Pearson graduated from Cambridge University in 1879 [the same year Einstein was born], ranking high in Mathematics. He then traveled to Germany to study at the University of Heidelberg, studying both physics and metaphysics. He returned to England in 1880, going to Cambridge and for a while studying Law. During 1882-84 he lectured around London on a wide variety of topics, also writing essays, articles, and reviews as well as substituting for professors of mathematics at KC and UCL, leading to him appointed to the Chair of Applied Mathematics in 1885.

Pearson was highly productive during the next 5-10 years. One book he completed was The Grammar of Science (1892), which prior to World War I was considered one of the great books on the nature of science and math, with perhaps the primary reason being it accessibility to practically anyone.(Another reason was its anticipating some of the ideas of Einstein’s Theory of Relativity.)
 
Pearson was married in 1890, with son Egon Pearson, who was to become another big contributor to Statistics, born in 1985.
Pearson was heavily influenced by Galton’s 1889 book Natural Inheritance, becoming interested in developing mathematical models for studying the processes of heredity and evolution. His 18 papers entitled Mathematical Contributions to the Theory of Evolution, written from 1893 to 1912, contain his most valuable work, including contributions to correlation and regression analysis and including the Chi-Square Test (1900), thus making it the oldest inference procedure still in use. (It has been said the chi-square test was produced in an attempt to remove the normal distribution from its central position.)

In 1900, Pearson, with W.F. Raphael Weldon and Galton, in response to biologists at the time not accepting biological conclusions based on mathematical analysis, founded the journal Biometrika, with Pearson coining its name and being its first Editor from the first issue, which appeared in October 1901, until his death almost 35 years later.
  
In 1911, Pearson started the Department of Applied Statistics at UCL, which was the result of Galton combining two other units to form it.

With World War I beginning in 1914, another sort of war had not yet begun. In September 1914, Pearson received a manuscript from a 24-year-old person by the name Ronald Fisher. The two corresponded in a friendly manner during the next couple of years until May 1917 when Pearson and some of staff criticized it unfairly, misunderstanding assumptions of Fisher’s maximum-likelihood method. By this time Fisher sensed that Pearson had been responsible for the rejection of many of his earlier papers, and the Pearson-Fisher War was on. (Pearson stopped considering any paper by Fisher to appear in Biometrika. As a matter of fact, the Fisher’s “maximum likelihood” didn’t appear in print (elsewhere) until 1922. Also, Fisher turned down an offer as Chief Statistician in 1919 because he would have to work under Pearson. The dispute between them continued to grow in intensity year after year, with each taking every opportunity to attack the other’s views.) (*** see addendum at the bottom)

Pearson became a Fellow of the Royal Society in 1896. His son Egon became head of a split Department of Statistics at UCL in 1933.
The following are many of the terms in Statistics due to Karl Pearson (along with the year first used, arranged chronologically):

moment (1893)
standard deviation (and use of sigma as, 1894) mode (1895) skewness (1895) correlation coefficient (1896; note: Galton first used "correlation" in 1888) chi-squared (1900) goodness-of-fit (1900) homoscedastic (1905) multiple correlation (1908)

 
For Pearson's connection to Einstein and much more, read the bio at:



Thanks Dave
*** A short while after I posted this, I came across a historical note about the Fisher-Pearson disagreements on a post about the Chi-square tests from Dan Teague of the North Carolina School of Science and Mathematics...

No comments: