STATISTICS

Bayesian

Critics of the Bayesian approach regarded its exponents as being a semi-religious movement. Dennis Lindley (1923-2014), an ardent exponent of the creed, was appointed to be the Professor of Statistics at University College, which until then had been a stronghold of the prevalent frequentist statistical orthodoxy. This prompted Patt Rivett, one of his colleagues, to exclaim It's as though a Jehovah's Witness has been elected Pope. Lindley had not intended to become a Bayesian. Rather, he had sought to furnish frequentist statistics with a system of axioms from which consequences could be deduced thereby conferring upon the same classical form that the other branches of mathematics had.

It was Lindley who coined Cromwell s Rule. This states that an assertion should not declare something either to be or impossible unless it is logically so. The name derived from Oliver Cromwell's 1650 letter to the Synod of the Church of Scotland. In this, the Protector wrote I beseech you, in the bowels of Christ, think it possible that you may be mistaken.

Francis Galton

The statistician and eugenicist Francis Galton's appreciation of regression and correlation did much to create the basis of statistics.

In 1906 Galton chose to spend a day in the countryside. He went to a county fair. There, he saw a competition that involved people estimating the weight of an ox. He took possession of the slips of paper upon which the contestants had written their estimates. He calculated their mean. It proved to be 1087 lbs.. The animal's actual weight was 1088 lbs.. No individual got the right number. While it went against his personal inclination towards elitism, he published his observation.

Location: University College, Gower Street, WC1E 6BT (purple, red)

John Graunt

In the 1660s John Graunt (1620-1674) created the first life tables.

Location: Birchin Lane, EC3V 9BW. Graunt's home. (blue, yellow)

Guinness

Guinness recruited a group of six extremely able scientists to improve the quality of the beer it made. Within the brewery they were treated with the utmost respect. It came to be appreciated that a reliable technique needed to be developed for using small samples of barley and hops in order to assess the probable quality of the overall crop. However, it proved that none of the young men was partial to doing the mathematics that would be required to develop a reliable statistical method. Finally, one of the group, William Sealy Gosset (1876-1937), agreed to undertake the work. In 1906 he travelled to London, where he studied with Karl Pearson (1857-1936) at University College. They developed the T Distribution. Using it, they identified a particular strain of barley as being optimum for brewing stout. The company used this information and its financial clout to establish a monopoly of the variety.

In 1911 University College established the world's first academic statistics department.

Location: University College, Gower Street, WC1E 6BT (purple, red)

Website: www.guinness.com

David Backhouse 2024