18
$\begingroup$

The history of many fields of science can be divided into a small number of time intervals that often begin with some important discovery.

But I have never seen something similar in timeline of statistics.

Obviously, there are some important dates that can be considered as starting points of a new period (Pascal+Fermat, Bayes, Pearson, Tukey,..).

Can we at least very roughly divide history of statistics into small number of periods? Note that the only similar question to this is related to only famous statisticians, not to periods in history.

$\endgroup$
  • 2
    $\begingroup$ Stephen Senn (amazon.com/Statistical-Issues-Development-Statistics-Practice/…) writes "A Brief and Superficial History of Statistics for Drug Developers" in its Chapter 2, which might be of interest... $\endgroup$ – ocram Oct 11 '12 at 7:41
  • 2
    $\begingroup$ Related, but not directly answering the question, Michael Friendly has written an article suggesting different periods in statistical graphics (Friendly, 2008). Also for data visualization Howard Wainer has some related work in which he suggests different periods (Wainer, 2001). $\endgroup$ – Andy W Oct 11 '12 at 13:01
  • 1
    $\begingroup$ There are a number of books on the history of statistics and some probably divide it in the way you would like. Aside from what has already been recommended consider Stephen Stigler's book and the book by Hald. The Kotz and Johnson three volume series on breakthroughs in statistics may also help point to significant papers that mark the beginning of a new point. You may find that periods should be shorter now because of the advances in computing. The 1980s introduced the idea of computer-intensive methods after Efron's 1979 paper. $\endgroup$ – Michael Chernick Oct 11 '12 at 20:10
  • 2
    $\begingroup$ MCMC started the Bayesian revolution of the 1990s with the rediscovery the Metropolis-Hastings algorithm/Gibbs Sampling. $\endgroup$ – Michael Chernick Oct 11 '12 at 20:11
  • $\begingroup$ I don't think a case can be made for Rev. Bayes marking any sort of milestone in Statistics (or Probability, for that matter). Bayes' Theorem was not published until after his death, and even at that is a fairly obvious corollary of the Total Probability Theorem. Laplace could be considered a proto-Bayesian, but I'd consider de Finetti, Jeffreys and maybe Savage to mark the real beginnings of Bayesian inference. $\endgroup$ – Dennis Jul 3 '14 at 4:47
18
$\begingroup$

These recent papers by Stigler, where he argues (convincingly I believe) for the types of periods you seem to have in mind.

Stigler, Stephen M. 2010. Darwin, Galton and the statistical enlightenment. Journal of the Royal Statistical Society: Series A 173(3):469-482.

Stigler, Stephen M. 2012. Studies in the history of probability and statistics, L: Karl Pearson and the Rule of Three. Biometrika 99(1): 1-14.

$\endgroup$
  • 3
    $\begingroup$ Welcome to our site! A high-voted answer is a great way to start out: I hope we will be seeing many more answers like this from you. $\endgroup$ – whuber Oct 12 '12 at 15:05
2
$\begingroup$

I think that "periods" in history are closely related to people and their developments. Of course one can expect "waves" in Toffler's sense, but even those waves are related to persons.

Anyway, wikipedia has an article in this regard.

$\endgroup$
  • $\begingroup$ Thanks for answer but this is only timeline. Imagine that you start any course of statistics and you first want to introduce to students some basic time periods of statistics. Of course, only very roughly. $\endgroup$ – Miroslav Sabo Oct 11 '12 at 7:55
  • 1
    $\begingroup$ I completeley agree, probability can be the first time period in history of statistics. But then, when the second period starts? :-) $\endgroup$ – Miroslav Sabo Oct 11 '12 at 8:14
  • 1
    $\begingroup$ I'm just checking this book and it seems very well explained. So at least you'll have the history before 1750. Then, you can complement with this other book, but I didn't check it yet. $\endgroup$ – Diego Oct 11 '12 at 8:26
  • 1
    $\begingroup$ sort of 1713-1935 $\endgroup$ – Diego Oct 11 '12 at 8:43
  • 2
    $\begingroup$ Another good book is The Science of Conjecture: Evidence and Probability before Pascal. Also The Unfinished Game $\endgroup$ – Peter Flom - Reinstate Monica Oct 11 '12 at 10:07
1
$\begingroup$

According to the webpage titled "Materials for the History of Statistics" by the Department of Mathematics at the University of York, a major text on this subject is:

Oscar Sheynin, Theory of Probability: A Historical Essay (published by NG Verlag 2005, ISBN 3-938417-15-3)

The book is packed full of names, dates, ideas, and references. It's probably a good contender for what you're looking for.

In the Preface to the book, the author tells us that:

The book is intended for those interested in the history of mathematics or statistics and more or less acquainted with the latter. It will also be useful for statisticians.

He then goes on to give a short outline of the book:

I describe the origin of the notions of randomness and subjective or logical probability in antiquity, discuss how laymen comprehended the main notions of the theory of probability, dwell on the birth of political arithmetic and study the history of the theory proper. I also trace the development of statistics and its penetration into natural sciences as well as the history of the mathematical treatment of observations (Ptolemy, Al-Biruni, Kepler, the classical error theory). I stop at the axiomatization of probability and at the birth of the real mathematical statistics, i.e., at Kolmogorov and Fisher.

The author appears to be active at making revisions to the book, so it would be worth visiting his website to see the latest available version of the book and his other related publications.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.