Stats without Tears
Big Names in Statistics

Dad sighed. “Kip, do you think that table was brought down from on high by an archangel?”

Robert A. Heinlein, in Have Space Suit—Will Travel (1958)

Math and science books tend to spend all their time teaching you the concepts, but the fact that these concepts were invented by people gets lost. Here’s a rundown of the people who invented most of the concepts that you meet in a first course in statistics. (I’ve made extensive use of Upton & Cook 2008 [see “Sources Used” at end of book] in preparing this page.)

Don’t panic! While I think it’s nice if you know that these people existed, you won’t be quizzed on who invented what.

See also: You might like to visit Figures from the History of Probability and Statistics (Aldrich 2012 [see “Sources Used” at end of book]). Aldrich puts these people, and lots more, in a time sequence and shows pictures of many of them.

Bernoulli, Jacob or Jacques (1654–1705), Swiss

Member of a famous mathematical family. Formulated the law of large numbers in 1689. Bernoulli trials are named after him, and he developed the binomial distribution, though this work was not published till eight years after his death.

Bradford Hill, Sir Austin (1897–1991), English

With Sir Richard Doll, published the first paper linking smoking and lung cancer, Smoking and Carcinoma of the Lung (1950). Is best known for the 1965 paper The Environment and Disease: Association or Causation? In it, he laid out nine criteria for confirming when A really is a cause of B and not merely associated. For a short summary of the criteria, see Wikipedia’s Bradford Hill criteria or Steve Simon’s Causation.

Fisher, Sir Ronald Aylmer (1890–1962), English

Referred to as R. A. Fisher. Coined the term variance and assigned it the symbol σ² in 1918. Developed the ANOVA procedure for testing equality of three or more means in 1918, and the F distribution in 1922. Published the Fisher z distribution (different from the normal distribution) in 1924 in a paper called “On a Distribution Yielding the Error Functions of Several Well-Known Statistics”; among other uses, it determines the confidence interval for the correlation coefficient of a population. Gave the mathematical derivation of Gosset’s t distribution in 1925. “Virtually invented the subject of experimental design” (Upton & Cook) and brought out editions of Statistical Methods for Research Workers every few years from 1925 until his death.

Gauss, Johann Carl Friedrich (1777–1855), German

Hugely prolific mathematician and astronomer. Said by Isaac Asimov to be the last mathematician to publish papers in Latin. Published in 1823 a paper with the resounding title “Theoria Combinationis Observationum Erroribus Minimis Obnoxiae”, giving the theory of least squares regression. As part of this work, he had to find an appropriate probability distribution for the errors, which we now know as the normal distribution.

Gosset, William Sealy (1876–1937), English

In 1908, while working for Guinness, he tested small samples of the product and realized that existing statistical theory of small samples was wrong. He then discovered and named the t distribution, which is followed by the means of samples where the standard deviation of the population is unknown. (R. A. Fisher later derived that distribution mathematically.) Because company policy did not allow him to publish under his own name, he used the pseudonym “Student”, and the t distribution is still known as “Student’s t” because of this.

Laplace, Pierre-Simon, Marquis de (1749 1827), French

Mathematician and mathematical physicist. Ennobled as a count by Napoleon in 1806, then created a marquis by Louis XVIII in 1817. Developed the Central Limit Theorem in Théorie analytique des probabilités (1812).

Neyman, Jerzy (1894–1981), Polish American

Studied under Karl Pearson. With Egon Pearson, published the 1933 paper that set out the standard method of testing hypotheses.

Pearson, Egon Sharpe (1895–1980), English

Son of Karl Pearson. With Jerzy Neyman, invented the standard approach to hypothesis testing, with null and alternative hypotheses. They published their paper, “On the Problem of the Most Efficient Tests of Statistical Hypotheses”, in Philosophical Transactions of the Royal Society in 1933.

Pearson, Karl (1857–1936), English

Invented the term histogram in lectures some time before 1895. Named the standard deviation in 1893, and gave the population standard deviation the symbol σ in 1894. Defined the correlation coefficient in a paper published in 1896. Devised the χ² goodness-of-fit test in 1900.

Tukey, John Wilder (1915–2000), American

Coined the word “bit” in 1946 and the word “software” in 1958. Introduced the stem-and-leaf plot and the box-whisker diagram in his book Exploratory Data Analysis in 1970. Invented the Honestly Significant Difference test to be done after an ANOVA. He wrote this up but circulated it informally from 1953 till it was finally published in J. W. Tukey’s Collected Works during the middle 1980s and early 1990s.