Home > Positioning > Persons > Pearson

Karl Pearson (1857–1936)

Pearson formalised the statistical methods that Galton had sketched and built the institutional apparatus that carried them forward. The product-moment correlation coefficient, the chi-squared test, the method of moments, the system of frequency curves — these became the standard toolkit of empirical science across disciplines. He founded the world’s first university statistics department, the journal Biometrika, and directed the Galton Laboratory for National Eugenics at University College London for three decades. The biometric programme he led — the study of heredity through the statistical relationships among measurable traits in populations — was both a productive research tradition and a vehicle for eugenic ideology. As with Galton and Fisher after him, the statistical innovations and the eugenic commitments are not separable: the methods were developed to answer eugenic questions, and the institutional home was a eugenics laboratory.

Life

Born 27 March 1857 in London. Originally named Carl, he changed the spelling to Karl in his twenties — reportedly in admiration of Karl Marx, though the political commitment did not last. Educated at University College School, London, then King’s College, Cambridge (Third Wrangler in the Mathematical Tripos, 1879). Studied law at Lincoln’s Inn and was called to the bar, but never practised. Studied physics and metaphysics at the universities of Berlin and Heidelberg. The breadth of his early interests — mathematics, philosophy, literature, law, German intellectual culture — shaped a career that was always more than narrowly statistical.

Appointed professor of applied mathematics at University College London (1884). His early published work included studies of medieval German literature, a philosophical treatise (The Ethic of Freethought, 1888), and The Grammar of Science (1892) — a philosophy-of-science work that influenced the young Einstein and anticipated logical positivism in its insistence that science describes correlations among sense impressions, not underlying causes.

The turn to biometry came through Galton. Pearson attended Galton’s lectures in the late 1880s and recognised that Galton’s insights about heredity and variation could be given rigorous mathematical form. The collaboration defined the rest of Pearson’s career. He developed the statistical methods; Galton supplied the biological programme and the institutional endowment. After Galton’s death in 1911, Pearson became the first Galton Professor of Eugenics at UCL and directed the Galton Laboratory until his retirement in 1933, when Fisher succeeded him — though the two men were personal and intellectual antagonists.

Fellow of the Royal Society (1896). Darwin Medal (1898). Founded Biometrika in 1901, with W. F. R. Weldon and Galton, as a journal for the statistical study of biological problems — partly in response to the Royal Society’s refusal to publish their longer statistical papers. Died 27 April 1936 in Coldharbour, Surrey.

Statistical methods

Pearson’s statistical contributions are foundational. They predate the modern concept of mathematical statistics as a discipline and were developed in the context of biological problems — heredity, variation, classification — rather than abstract mathematics.

The product-moment correlation coefficient (r). Galton had the concept of correlation; Pearson gave it its mathematical form. The Pearson correlation coefficient measures the linear relationship between two variables and remains the most widely used measure of association in the sciences. Pearson also developed the theory of multiple and partial correlation — the relationships among more than two variables, and the correlation between two variables after controlling for a third.

The chi-squared test (1900). A test for whether observed frequencies differ significantly from expected frequencies — applicable to contingency tables, goodness-of-fit tests, and tests of independence. The chi-squared test became one of the most widely used statistical tests in the sciences and remains so.

Method of moments. A general technique for estimating the parameters of a probability distribution from sample data, by equating sample moments (mean, variance, skewness, kurtosis) with theoretical moments of the distribution. Superseded in many applications by Fisher’s maximum likelihood method, but historically foundational.

The system of frequency curves. Pearson developed a family of probability distributions (the Pearson system) capable of fitting a wide range of empirical data — skewed, leptokurtic, platykurtic — providing flexible models for the distributions encountered in biological and social data.

The biometric programme and the Mendelian conflict

Pearson and Weldon developed the biometric approach to heredity: study the transmission of traits through the statistical relationships among parents and offspring, without assuming a particular mechanism of inheritance. The approach was productive — it produced real knowledge about the patterns of heredity in natural populations — and it was explicitly anti-Mendelian. Pearson and Weldon argued that Mendel’s discrete hereditary factors could not account for the continuous variation observed in most traits.

The conflict with the Mendelians — led by William Bateson, who coined the term “genetics” — was bitter and personal. Pearson used Biometrika as an institutional weapon; Bateson used the newly founded genetics journals. The dispute was partly scientific (continuous vs. discrete heredity), partly methodological (statistics vs. breeding experiments), and partly personal (Pearson and Bateson detested each other). The resolution came from Fisher, who showed in 1918 that continuous variation is compatible with Mendelian inheritance when many genes of small effect are involved. The reconciliation vindicated both sides: the biometricians were right that variation is continuous; the Mendelians were right that inheritance is particulate. Fisher’s synthesis left Pearson’s methods intact while undercutting his theoretical framework.

Eugenics

Pearson was a committed eugenicist throughout his career. He directed the Galton Laboratory for National Eugenics from 1911 to 1933. His eugenics was not incidental to his statistics — the laboratory’s research programme was explicitly eugenic, and the statistical methods were developed and applied in the service of measuring human variation, classifying it, and drawing conclusions about which populations were “fit” and which were not.

Pearson’s eugenic work included studies of the inheritance of intelligence, physical fitness, and disease, as well as comparative studies of racial and ethnic groups that drew conclusions about the relative “quality” of different populations. His 1925 paper “The Problem of Alien Immigration into Great Britain, Illustrated by an Examination of Russian and Polish Jewish Children” used biometric methods to argue against Jewish immigration — work that was eugenic in motivation and racist in its conclusions. The paper is representative of a broader pattern in Pearson’s applied work: the statistical methods are technically competent; the questions they are asked to answer are shaped by eugenic and racial ideology.

The institutional legacy is direct. The Galton Laboratory, under Pearson and then Fisher, was the centre of British eugenics and the birthplace of population genetics. University College London’s 2020 review of its eugenics history examined the institutional context in which Pearson and Fisher worked and renamed buildings and professorships that had carried their names.

Where Pearson stops

Pearson’s biometric programme studied heredity at the phenotypic level — statistical relationships among measurable traits — without a mechanism of inheritance. This was a methodological choice, not an oversight: Pearson believed that science describes correlations, not causes (The Grammar of Science made this explicit). But the choice had a cost. When Fisher demonstrated in 1918 that Mendelian genetics produces exactly the continuous variation the biometricians observed, Pearson’s theoretical framework was superseded even as his statistical methods remained indispensable. The methods outlived the programme that produced them — a pattern Pearson shares with Galton.

The Pearson-Fisher antagonism was personal and institutional but also methodological. Fisher’s maximum likelihood estimation and his approach to experimental design (randomisation, replication, factorial design) displaced Pearson’s method of moments and his preference for observational data. Fisher argued that Pearson’s methods were inefficient — that they discarded information available in the data — and that Pearson’s reluctance to adopt Mendelian genetics was not methodological scruple but intellectual stubbornness. The dispute was not resolved in Pearson’s lifetime; Fisher’s methods became dominant after his death.

Pearson’s philosophy of science — the strict positivism of The Grammar of Science — was influential in its time (it shaped logical positivism and influenced Einstein’s early epistemology) but proved too restrictive for the biology Pearson himself practised. His formulation was uncompromising: “The unity of all science consists alone in its method, not in its material. The man who classifies facts of any kind whatever, who sees their mutual relation and describes their sequences, is applying the scientific method and is a man of science.” Causal language, on this view, is metaphysical overreach; science describes sequences and correlations, not underlying mechanisms. The biometric programme needed to say something about heredity, not just about correlations among measurements. Pearson’s positivism prevented him from engaging with the mechanistic question that Fisher and the geneticists answered.

Key works

See also: Galton · Fisher · Mendel · Darwinism