THE EXAMINED LIFE:
COMPETITIVE EXAMINATIONS IN THE THOUGHT OF FRANCIS GALTON
Raymond E. Fancher
"I look upon social and professional life as a continuous examination," wrote Francis Galton in Hereditary Genius (Galton, 1869, p. 49). In expressing this belief Galton was very much a man of his time, for Victorian England was marked by great enthusiasm for the institution of competitive examination, as means of allocating rewards in the universities, the schools, and the public service (Roach, 1971). Accordingly, it is not surprising to find that Galton's pioneering work on mental testing - including the famous South Kensington Anthropometric Laboratory whose centenary we observe this year - was always closely associated in his mind with the basic model of competitive academic examinations.
Even as a four-year-old child Francis Galton's fondest wish was to win honors at the university (Pearson, 1914, p. 69n), and this remained his major goal when he entered Cambridge University in 1840. During his pre-university career he had shown a keen interest in academic competition (Fancher, 1983) and at Cambridge he naturally became preoccupied with the only vehicle to an honors degree, the Mathematical Tripos examination. This test - dating from the early 1700s and the oldest competitive examination in England - precisely ranked the 100 or so honors-winning graduates of each class from first to last, and sorted them into three categories: the top 35 or 40 were designated "Wranglers," while the remainder were called Senior or Junior "Optimes." An additional 300-350 students each year either failed the Tripos, or elected not to compete for honors at all. The Tripos results were always published and aroused great general interest even among the college servants, who often laid wagers as to who would finish as the top or Senior Wrangler (Roach, 1971, pp. 8-14).
During his early terms at Cambridge, Galton developed a keen critical interest in the general techniques of academic examining. Thus he complained to his father about some first year College tests he had taken: "The Math. papers were exceedingly easy this year so that everybody who knows anything about them must of necessity do three-fourths, hence there was little room for a man to distinguish himself in them'' (letter of 11 June 1841, quoted in Pearson, 1914, p. 154). Much more to his taste was the Tripos for the class ahead of his, in which, he wrote his father, "Cayley is 1st and Simpson 2nd Wranglers. These were very far superior to the rest. [My tutor] told me today that Simpson was 1000 marks ahead of the 3rd Wrangler, and the getting of 500 marks only entitles a man to be a Wrangler" (letter of 21 January 1842, quoted in Pearson, 1914, p. 164). These were typical of Cambridge Tripos results, and 27 years later in Hereditary Genius, Galton would cite the extreme dispersion of the top Wranglers' scores as evidence for the normal distribution of mathematical ability - an important part of his argument for the hereditary determination of intellectual ability (Galton, 1869, pp. S8ff.). It seems likely the idea was first impressed upon him by these 1842 results.
Galton did well enough in his own early examinations to sustain hopes for an eventual Wranglership, but in March 1842 he suffered severe disappointment in the "Little Go," the most important university-wide examination before the Tripos. He finished ih the second class - a creditable enough performance for many people, but shattering for Galton. He promptly withdrew from a forthcoming scholarship examination, on grounds that he already knew his place and had nd chance of winning, and then suffered a disabling emotional breakdown marked by obsessional thoughts and dizziness whenever he tried to study mathematics. The next year he withdrew from the Tripos completely, and settled for a non-honors or "poll" degree. Galton's Little Go disappointment could not be blamed on external factors, and so convinced him of his own innate mathematical limitations and helped produce his strongly nativistic psychology (Fancher, 1983).
The theme of examinations recurred strikingly in Galton's thoughts as he wrote "Hereditary Talent and Character," the pivotal 1865 paper characterized by Pearson as "an epitome of the great bulk of Galton's work for the rest of his life" (1924, p. 86). Among many other contributions, Galton here first introduced the basic idea of the intelligence test, as a means of selecting highly able young men and women for eugenic breeding purposes. He asked his readers to imagine some future Utopia, "in which a system of competitive examinations for girls, as well as for youths, had been so developed as to embrace every important quality of mind and body, and where a considerable sum was allotted [by the State] to the endowment of such marriages as promised to yield children who would grow into eminent servants of the State" (Galton, 1865, p. 165). The top-scoring young men and women on each year's examinations would be presented at a public ceremony, introduced to each other, and offered financial inducements to intermarry and have many children. Here were honors examinations with a new twist!
Of course, it was one thing to imagine the future existence of such examinations, and quite another again to develop them in reality. Galton's best known approach to this practical problem was demonstrated in his Anthropometric Laboratory of 1884, where he measured head size, reaction time, and several varieties of sensory acuity, in hopes that these would reflect innate neurophysiological vigor and efficiency, and hence intelligence or mental ability as well. Galton himself did only a rough statistical analysis of these test results, but his Anthropometric Laboratory directly stimulated James McKeen Cattell's influential (if ultimately disappointing) work on "mental testing" throughought the 1890s (see Cattell, 1890; Sokal, 1981).
Even as he was developing this physiologically-based, "anthropometric" approach to testing, Galton believed that more traditional examination procedures might prove useful for his eugenic purposes as well. He wanted tests which could be given to young adults, and could accurately predict their intellectual achievements in full maturity. An excellent potential motel here seemed to be the competitive examinations required of all candidates for the Indian Civil Service since the mid-1850s. Open only to men in their early 20s, these examlnations covered English, mathematics, classics, modern languages, and science, as well as Indian history and society. Modelled explicitly on Oxford and Cambridge examinations and characterized as "the honours examinations of the public service" (Roach, 1971, p. 195), these were among the first competitive examinations in Britain to be used outside a strictly academic setting.
As part of his eugenic testing program, Galton hoped to show that the civil service examinations given in the early 1860s accurately predicted later success in the Indian Service. Thus in 1883 he obtained the original examination scores of some 250 men who had entered the Indian Civil Service between 1861 and 1865, together with their salary figures for 1882. Essentially, he hoped to show a strong positive correlation between the two variables - though at the time he had not yet developed the epoch-making approach to correlation he would introduce in 1888 (Galton, 1888). Thus his analysis of the Indian Civil Service data (contained in a handwritten manuscript currently held in Folder #137/9 of the Galton Papers, housed in the Library at University College London) provides an interesting view of his early thinking on the subject of correlation and regression, as well as on examinations in general.
Galton first showed that out of the 25 top men in examination results (the first 5 for each of the years 1861 to 1865), 3.5 (one case involved a tie) occupied exactly the same rank, or the one immediately above or below it, in the 1882 salaries list. He calculated that the odds against getting as many as 3 "hits" of this type by chance were more than 20 to 1. For the bottom 25 exam scorers he found only 2 such hits, however, a result which would be expected by chance alone about a quarter of the time. Galton apparently did not analyze the middle portions of the exam lists in this manner.
In a different approach, Galton selected the top five, middle five, and bottom five examination scorers for each of the five years, and then calculated the mean examination scores and 1882 salary levels separately for all the groups. He then arbitrarily set the exam and salary means for the middle group at 100, and adjusted the means of the top and bottom groups proportionately. When he did so, he found that the examination scores for the top and bottom groups differed as 133.6 to 83.8, while their salaries differed only as 115 to 103. (Thus the bottom group of examination scorers were earning some 3 per cent more than the middle group, after 20 years.) Galton summarized these findings by writing, "The lowest class men are receiving 10 per cent less pay after 20 years service, though their (examination) marks were 40 per cent less." Thus Galton was already thinking along the lines of regression, and it is interesting to note that his observed ratio of 10 per cent to 40 per cent closely approximates the mean Pearson r of .26 which I found after re-analyzing his original data with modern techniques.
Galton's correlational ideas are commonly believed to have originated in his investigations of parent-child similarities, or the interrelations between physical measures such as height and arm-length (Forrest, 1974, pp. 196 ff.). The Indian Civil Service manuscript shows that his studies of examination validity may have played a role as well. [For a later and more detailed analysis and discussion of Galton's Civil Service study see Fancher, 1989.]
Galton never published these rather weak findings, which must have disappointed him, but he never gave up hope that the examinations would prove more valid in the long run. Eighteen years later he wrote Carl Pearson: "I have [an investigation] in view now, that I began upon some years ago, but found that enough years had not elapsed since the experiment began to draw useful conclusions . It is the correlation in the Indian Civil Service between the examination place of the candidate and the value of the appointment held by him afterwards" (letter of 25 October 1901, in Pearson, 1930, p. 247). He hoped that by 1901, with his men in their 60s and at the ends of their careers, he could get more accurate measures of their overall success, and that these would correlate more strongly with examination scores than his figures of 1882 did.
I have found no evidence that the aging Galton ever collected data for this further study, but he did unquestionably maintain his faith in the potential eugenic value of honors-type examinations. Pearson (1930, pp. 411 ff.) has published the surviving fragments of Kantsaywhere, a short utopian novel Galton wrote in 1910, just before he died. In this fantasy, a young Englishman finds himself in "Kantsaywhere," a country which has instituted highly effective eugenic practices. Preeminent among these are two examinations administered by the local "Eugenic College." First is a largely medical "Pass Examination" which must be taken by everyone and passed to gain State permission to marry and have children. A sufficiently high pass on this entitles one to compete in the further "Honours Examination," with four equally weighted parts measuring a) medical fitness; b) ancestral quality; c) anthropometric quality; and d) aesthetic and literary skill. The anthropometric tests in this examination are exactly like those from the South Kensington Laboratory. The "aesthetic and literary" section includes assessment of aesthetic judgment, as well as more standard examination tasks such as essay writing.
Each of the four examination sections is worth 30 points, and the hero is informed that half of the final total scores each year fall between 45 and 70. The quarter or so of each year's candidates scoring above 70 get their names in the newspaper - just like Cambridge Optimes. The Kantsaywhere equivalents of Wranglers receive particular attention, because any couple amassing 200 or more points between themselves can marry and rear children with extremely generous State support.
Recalling Galton's own youthful experience at Cambridge, it is amusing to observe the fate of his fictional hero. Kantsaywhere's lovely heroine, Miss Augusta Allfancy, has already scored 94 on her Honours Examination, and the hero elects to compete himself in hopes of justifying a marraige proposal. After earning a first-class Pass degree, he gets 77 on the Honours Examination - a score which is probably artifically depressed because he cannot document the quality of his English ancestry and earns only 5 out of 30 possible points in that section of the examination. Augusta's father is impressed by this performance, which gets the hero's name in the newspapers, but must naturally consfider the possibility that another suitor will appear with enough marks to qualify for a state-supported marriage. Unfortunately, the outcome remains in doubt because the relevant parts of Kantsaywhere were seen by Galton's executors as absurd, and destroyed.
In sum, the image of competitive examinations, and of the poignant experiences of those who take them and just fall short of outstanding success, remained vivid in Galton's thought throughout his life. To the very end, he believed that anthropometric and more orthodox examination procedures might one day be combined to yield valid predictions of young people's eugenic worth. The overall model for this procedure was always the Cambridge Tripos, in the preparation for which he had himself undergone such emotionally significant experiences in his youth.
Note: Research for this paper was conducted while the author was on sabbatical leave at the History of Medicine Unit of University College London, with the generous support of a Leave Fellowship from the Social Sciences and Humanities Research Council of Canada.
Cattell, J. Mck. (1980). Mental tests and measurements. Mind, 15, 373-381.
Fancher, R. E. (1983). Biographical origins of Francis Galton's psychology. Isis, 74, 227-233.
Fancher, R. E. (1989). Galton on examinations: An unpublished step in the invention of correlation. Isis, 80, 446-455.
Forrest, D. W. (1974). Francis Galton: The life and work of a Victorian genius. London: Elek.
Galton, F. (1865). Hereditary talent and character. Macmillan's Magazine, 157-166, 318-327.
Galton, F. (1972). Hereditary genius. Gloucester, Mass.: Peter Smith, (originally published 1869)
Galton, F. (1888). Co-relations and their measurement, chiefly from anthropometric data. Proceedings of the Royal Society 45, 135-145.
Pearson, K. (1914-1930). The life, letters and labours of Francis Galton. Cambridge, UK: The University Press, (Vols. 1-3)
Roach, J. (1971). Public examinations in England 1850-1900. Cambridge, Eng.: The University Press.
Sokal, M. (1981). An education in psychology. Cambridge, Mass.: MIT Press.