Lecture 5

readings 2018-02-12

1. Desrosieres, Alain. "Correlation and the Realism of Causes," in The Politics of Large Numbers: A History of Statistical Reasoning. Cambridge, Mass.: Harvard University Press, 1998, ch 4.

2. Galton, Francis. “Typical Laws of Heredity,” Royal Institution of Great Britain. Notices of the Proceedings at the Meetings of the Members 8 (February 16, 1877): 282ff.

3. Stephen J. Gould, Mismeasure of Man, ch. 3

OPTIONAL

4. Student (W. S. Gosset), "The Probable Error of a Mean," Biometrika, 6 (1908), 1-25. (https://www.jstor.org/stable/2331554)

5. Gillham, Nicholas. "Sir Francis Galton and the Birth of Eugenics." Ann. Rev. Genet. 35 (2001): 83-101.

discussion:

1a:

• "statistics" meant the state

Opening paragraph of Des (was not assigned, but so good I must share):

What do statistics—a set of administrative routines needed to describe a state and its population; the calculus of probabilities—a subtle manner of guiding choices in case of uncertainty, conceived circa 1660 by Huygens and Pascal; and the estimates of physical and astronomical constants based on disparate empirical observations, carried out around 1750, all have in common? Only during the nineteenth century, after a series of retranslations of the tools and questions involved, did these various traditions intersect and then combine, through mutual exchanges between the techniques of administrative management, the human sciences (then known as “moral sciences”), and the natural sciences.

• "vulgar" implies contested quantification leitmotif

a particular country or individual person cannot be reduced to characteristics selected precisely in order to permit comparison.

• tables as example of "toolset->mindset" leitmotif

“It was thus the tabular form itself that prompted the quest for and comparison of numbers. This form literally created the equivalence space that led to quantitative statistics.” - Des

1b, on Galton, along with 2. Galton himself

• enamored of cousin's theory

“Evolution and quantification formed an unholy alliance; in a sense, their union forged the first powerful theory of "scientific" racism—if we define "science" as many do who misunderstand it most profoundly: as any claim apparently backed by copious numbers” -- gould p106

• obsessed with quantification as truthiness Contrast with Quetelet
• correlation and causation conflation [125]
• leitmotif: scientifically proving one's privilege
• Q-vs-G:

“This is where the essential difference arises between his con- struct and Quetelet’s. Galton cast his attention on the differences between individuals, on the variability of their attributes, and on what he would later define as natural aptitudes, whereas Quetelet was interested in the average man and not in the relative distribution of nonaverage men.” -des p113

• Q focus on reification of the average v G focus on ordering individuals and class
• G focus on the individual not the society; more technically: - normal curve as law of deviation not law of errors [113]

• isolation of causes vs. Quetelet -- development of regression [115]

3. Gould

• leitmotif: subjective design choices in data cleaning Foreshadows p-values as a defense against “Never has so much been coaxed from so little!”, p137
• leitmotif: scientifically proving one's privilege
• Gould's macro-message: IQ and g-factor (next week) and the general mismeasuring of man Leitmotif: Ranking (often) has no ground truth (later: search, USNWR):

“anthropometry became a search for characters that would display the correct ranking, not a numerical exercise in raw empiricism.” p118

2017

readings

1) Desrosières, Alain. "Correlation and the Realism of Causes," in The

Politics of Large Numbers: A History of Statistical Reasoning. Cambridge, Mass.: Harvard University Press, 1998, ch 4.

outline of Desrosières ch 4

1. (2pp) Intro
2. (7pp) Pearson
3. (9pp) Galton

• Box: "Social Classes and [sic] Genetic Worth"
• Box: Gould quote
• Box: Quincunx
• Box: Galton's regression curve as an elliptical joint distribution
• Box: 2-stage Quincunx

4. (6pp) Calculations
5. (8pp) 5 Englishmen
6. (9pp) Controversies on realism
7. (8pp) Yule and realism
8. (2pp) Spearman on "General Intelligence"

2) Galton, Francis. “Typical Laws of Heredity,” Royal Institution of Great

Britain. Notices of the Proceedings at the Meetings of the Members 8 (February 16, 1877): 282ff.

3) Stephen J. Gould, Mismeasure of Man, ch. 3

OPTIONAL

4) Student (W. S. Gosset), "The Probable Error of a Mean," Biometrika, 6

(1908), 1-25. (https://www.jstor.org/stable/2331554)

5) Gillham, Nicholas. "Sir Francis Galton and the Birth of Eugenics." Ann.

Rev. Genet. 35 (2001): 83-101.

discussion

broca: bad data practice

• logical fallacies vs bad stats/bad math
  • e.g., circularity -preview: "ranking" as illposed

bad blood / obscured+censored origins of statistics of statistics

• "regression"
• "co-relation"

from Quetelet to Galton

• the normal curve (and relations to "the bell curve")
  • as error
  • as ranking
• description+prescription (and preview: correlation v. causality)
  • preview: yule + multivariate regression
  • link: realism vs effective models
  • impact OF law (e.g., Yule is modeling Poor law of 1864 (?))
  • impact ON law
  • untouched: should society be dictaed science
  • preview: pearson the mathematizer

not discussed:

Quetelet

• science envy
• "policy" and "social physics"

Desrosières

• reality of society

• what we talk about when we talk about averages

  • Statue example: corrupt versions of a perfect reality
  • objective mean: measuring a single object with errors
  • subjective mean: average over a collection of members of a defined population
  • arithmetic mean: average of a feature without any grounds for positing a distribution (and without meaningful cause)

• causal story in the above: ``constant cause"

• statistics as measurements of morality

  • crime rates

Levels of analysis in Desrosières

• philosophy of science/epistemology
• mathematical technique
• ontology: what the world is made up of
• disciplines
  • sociology and social physics
  • public health
• social/political framing
  • debates on expertise
  • quantification of arts
    • medicine
    • moneyball
    • politics and politicking
• in history