Sir Francis Galton - 100 years on

Author: Graham Wheeler

Sir Francis Galton in the 1850s or 1860s.

Galton in the mid-19th century. 

Scanned from Karl Pearson's

'The Life, Letters, and Labors of Francis Galton'

and available via Wiki Commons 

Monday 17th January 2011 marked the 100th anniversary of the death of Sir Francis Galton FRS (Fellow of the Royal Society). A man who was undoubtedly one of the most prominent scientific figures of his time, his work across a plethora of fields – including anthropology, dermatoglyphics, geography, meteorology, psychometrics and statistics – has shaped and influenced much subsequent research via the provision of novel theoretical and practical tools.

For most, Galton is probably characterized by one, if not both of these descriptions: 1) a forefather of eugenics (he invented the term in 1883), and 2) the half-cousin of Charles Darwin. Despite his somewhat controversial theories and publications on the subject of eugenics and his familial connection to the pioneer of modern evolution theory, his most astounding and relevant work receives much less attention from the wider world. Especially in the field of statistics, Galton has provided much more than one may think.

Galton’s explorations in the fields of heredity and ancestry led him to writing about “regression toward the mean” - the statistical phenomenon that if an initial measurement of some variable is an extreme value, then a second measurement will on average be closer to the true mean. Galton first uncovered this concept when looking at the size of sweet peas and the seeds produced, later extrapolating to heights of persons and the deviations from the heights of that person’s parents [1]. Although Galton slightly misunderstood regression toward the mean (he failed to acknowledge random variations in height, and proposed that the heights of more distant ancestors had a role to play), he certainly helped shed some light on the matter and this helped pave the way for other exciting research.

Just a placeholder image

Extract of Galton’s work on sweet pea size and size of seeds produced. Image from galton.org

Galton studied the Normal distribution during the 1870s and 1880s, and subsequently invented the Quincunx, or Galton Board to demonstrate the Central Limit Theorem – that the average of a large collection of independent measurements of some variable (e.g. height, weight) will be Normally distributed (for an illustration of this, watch the following video). He has also been credited with the “discovery” of the standard deviation, which enabled him to understand the variability of numerous characteristics within a population.

Perhaps his greatest statistical legacy was far more indirect, via the supervision of a doctoral student at University College London called Karl Pearson. Galton’s initial work on correlation was adapted by Pearson, yielding the commonly used Pearson’s Correlation Coefficient. Pearson was Galton’s protégé and together they collaborated with Walter Weldon to establish Biometrika, a leading scientific journal on biometrics and the analysis of biological phenomenon using statistics.

Galton worked throughout his life in almost every academic discipline – and often provided fundamental contributions that are still evident today. On top of his work in statistics, he is also attributed with producing the first weather map to be published in The Times (April 1st 1875), identifying the existence of the anti-cyclone, devising a fingerprint-classification used in forensic science and even founded the field of psychometrics. A true polymath of his time, Galton’s contributions to the sciences should not be underestimated, nor overshadowed by some of their applications.

Sir Francis Galton. Born 16th February 1822. Died 17th January 1911.

References

Bookmark and Share

Comment on this article

Submit your comment
  1. Image of unique ID

Comments

M. Eileen Magnello

   Francis Galton never held a university post and, therefore, never supervised any doctoral students at UCL. Moreover, Karl Pearson did not do a PhD in statistics at UCL, especially as UCL was not even offering any courses or degrees in mathematical statistics when he was a student.  Pearson not only created the foundations to modern mathematical statistics at UCL beginning in 1892, but he set up the first known university undergraduate degree in statistics in the world in 1917. 

By the time Weldon introduced Galton to Pearson, he had already set up the infrastructure to his statistical methodology. Additionally,  it was W.F.R. Weldon who first suggested to Pearson that they establish Biometrika in November 1899: Galton was asked later to ask in consultation on the journal, but  virtually all of the editing from 1901 to 1933 was done by Pearson.  

Finally, Pearson was not Galton’s protégé; the inpetus to  Pearson’s change of careers from an elastician to that of a mathematical statistician came  from Weldon in the early 1890s. It was Pearson’s later exploration of the mathematical properties of Galtonian correlation that  enabled him to devise a battery of correlational techniques:  Galton’s  influence does not extend any further than this.   Pearson developed a body of methods in mathematical statistics that lay outside Galton’s conceptual framework and beyond his technical scope. He was never Galton’s student and he was his intellectual heir in a limited sense only. Three men may be viewed as having inherited Galton’s statistics to a more notable extent include W F R Weldon,  Francis Ysidro Edgeworth and George Darwin.

reply to this comment

Skip to Main Site Navigation / Login

Site Search Form

Site Search