\[ %% % Add your macros here; they'll be included in pdf and html output. %% \newcommand{\R}{\mathbb{R}} % reals \newcommand{\E}{\mathbb{E}} % expectation \renewcommand{\P}{\mathbb{P}} % probability \DeclareMathOperator{\logit}{logit} \DeclareMathOperator{\logistic}{logistic} \DeclareMathOperator{\SE}{SE} \DeclareMathOperator{\sd}{sd} \DeclareMathOperator{\var}{var} \DeclareMathOperator{\cov}{cov} \DeclareMathOperator{\cor}{cor} \DeclareMathOperator{\Normal}{Normal} \DeclareMathOperator{\MVN}{MVN} \DeclareMathOperator{\LogNormal}{logNormal} \DeclareMathOperator{\Poisson}{Poisson} \DeclareMathOperator{\Beta}{Beta} \DeclareMathOperator{\Binom}{Binomial} \DeclareMathOperator{\Gam}{Gamma} \DeclareMathOperator{\Exp}{Exponential} \DeclareMathOperator{\Cauchy}{Cauchy} \DeclareMathOperator{\Unif}{Unif} \DeclareMathOperator{\Dirichlet}{Dirichlet} \DeclareMathOperator{\Wishart}{Wishart} \DeclareMathOperator{\StudentsT}{StudentsT} \DeclareMathOperator{\Weibull}{Weibull} \newcommand{\given}{\;\vert\;} \]

Eugenics and statistics

Peter Ralph

Advanced Biological Statistics

Some history

Threads of statistical history

The “birth of modern statistics”

Galton and Pearson RA Fisher

Galton’s science

plot of chunk r plot_galton

Karl Pearson and the Biometrics Laboratory

  • developed the correlation coefficient and the chi-squared test; early use of the \(p\)-value

  • head of the first Department of Applied Statistics, as the Galton chair of Eugenics

  • focus on “biometry”: statistical theory of biological measurements, especially evolution and inheritance.

  • had a lengthy feud with Udny Yule over the correct definition of the correlation coefficient for a \(2\times 2\) table

Biometrika cover

RA Fisher

  • developed ANOVA, F-distribution, methods for maximum likelihood, lots of evolutionary theory

  • Galton chair of Eugenics, succeeding K. Pearson

Statistical Methods for Research Workers 1932 cover

Some natural questions to ask

  1. What’s up with this connection to eugenics?

  2. How can we reconcile these people being very smart about some things but not about others?

  3. Has this history had an impact on how we do statistics?

Eugenics?

Now, if you are going to take Darwinism as your theory of life and apply it to human problems, you must not only believe it to be true, but you must set to, and demonstrate that it actually applies. […] It was not a light task, but it gave for many years the raison d’etre of my statistical work.

Pearson, 1912, in ‘Darwinism, Medical Progress and Eugenics’. Eugenics Laboratory Lecture Series, IX

Eugenics poster

Who gets to “direct”?

Galton’s view of British social structure

Galton 1909, Essays in Eugenics. London: Eugenics Education Society. (from MacKenzie

  • upper-class Nordic eugenicists thought that everyone except upper-class Nordic people were dangerously degenerate

  • they opposed curing tuberculosis and justified the genocide of the Americas

  • this inspired forced sterilization and anti-miscengenation laws in the US, targeting poor and people of color

  • and this directly inspired the Nazis

She has her mother’s laugh

PSA: eugenics is a bunch of bullshit

Others

Bystanders:

  • William Gosset (\(t\)-test): focused on practical applications

  • Egon Pearson (hypothesis testing): pushed to separate statistics and eugenics departments

Opponents:

  • Lancelot Hogben: biologist, opponent of eugenics

  • Thomas Morgan: biologist, opponent of eugenics

“Smart” people

Pearson and tuberculosis

  • caused by Mycobacterium tuberculosis

  • occurs more often among relatives

  • major risk factors: silicosis (30x), malnutrition, smoking, alcoholism, crowding

  • minor: genetics (mostly immune-related)

Pearson: it’s hereditary!

The hereditary factor in pellagra
  • niacin deficiency

  • was common among poor Americans in the early 1900s

  • … since corn lacks niacin unless nixtamalized.

Davenport: it’s hereditary!

Fisher and smoking

  • Hill & Doll: retrospectively compare cancer rates between matched smoking/nonsmoking pairs

  • Hill & Doll: prospectively compare cancer rates between smoking/nonsmoking doctors

  • etcetera

  • Fisher: Correlation is not causation. Let’s calmly study this a while longer.

fisher, smoking

what is being “smart”, anyways?

Has this affected how we do statistics?

well, numbers aren’t racist?

but how we use them might be

So the argument goes:

\(p\)-values and hypothesis tests focus on whether there’s a difference between groups,

… ignoring whether any difference is important.

This is a major problem with reflexive reporting of “statistical significance”.

Is this because of “eugenics thinking”?

A small methodological fix:

the confidence interval

(introduced by Jerzy Neyman in 1937)

More important than statistical technique:

  • What questions are being asked?

  • What data is being collected? (and how)

  • What assumptions are being made?

  • What are the important conclusions?

In conclusion

Looking forward

Pearson & Fisher had a big impact in part because (they were pushy and) they were working towards something.

Today we have many problems that we need quantitative work to solve.

Where do you want to have an impact?

David Blackwell FN David Bin Yu Simone Gray

Some sources

National Life, p.19 National Life, p.22

Karl Pearson, National Life from the Standpoint of Science, 1900.

It is a false view of human solidarity, a weak humanitarianism, not a true humanism, which regrets that a capable and stalwart race of white men should replace a dark-skinned tribe which can neither utilize its land for the full benefit of mankind, nor contribute its quota to the common stock of human knowledge.

Karl Pearson, The Grammar of Science, 1892.

// reveal.js plugins