8
$\begingroup$

There is a well-known statistical example, claiming that there is correlation between the number of babies in Alsatian/Danish/Dutch/German villages or European countries and the number of storks in that place. The humorous "implication" of this example is that storks indeed bring babies. This example is often used to teach the difference between correlation and causality.

I would like to know by whom this example was initially coined. So far, I found two references talking about this example. I checked this answer and found

Sies, H. (1988), A new parameter for sex education, Nature 332, 495; https://doi.org/10.1038/332495a0

Furthermore, I found

Matthews, R. (2000), Storks Deliver Babies (p= 0.008). Teaching Statistics, 22: 36-38. https://doi.org/10.1111/1467-9639.00013

Across the web, I also found attributions to G. Udny Yule, author of Introduction to the Theory of Statistics (1911) and to Darrell Huff, author of How to Lie with Statistics (1954). In neither of the books, I could find a paragraph on the subject. As G. Udny Yule is the oldest of the named authors, I would like to believe that he is the inventor of the storks-and-babies example, but I would like to see a reference proving or falsifying my assumption.

$\endgroup$
4
  • $\begingroup$ Yule was British (Scottish born). Storks aren't often seen in Britain. I think the folklore (or tale told to children before they are deemed old enough to learn otherwise) that storks bring babies is central European, although I've been hearing that story mentioned in Britain since childhood, perhaps by way of Grimm or Andersen. My guess is that this arose on the continent, as British people often say. (No Brexit jokes, thanks; it's not funny.) $\endgroup$
    – Nick Cox
    Feb 7, 2021 at 10:56
  • $\begingroup$ Thank you for the comment. I would like to clarify that I am not looking for the initial reference of the storks-and-babies tale, but for the initial reference of the storks-and-babies statistical example. $\endgroup$
    – Lukas
    Feb 7, 2021 at 11:25
  • $\begingroup$ That was clear to me; in turn I will summarise my earlier comment this way: as storks bringing babies appears to be a central European meme, it is more likely to spring to mind to a statistically minded person in those countries. Usages likely to be used informally first -- e.g. in teaching as a humorous but provocative example -- are in my experience especially difficult to track down. $\endgroup$
    – Nick Cox
    Feb 7, 2021 at 11:37
  • $\begingroup$ Googling "storks bring babies legend" brings up some details. Andersen did re-tell the story. $\endgroup$
    – Nick Cox
    Feb 7, 2021 at 11:40

2 Answers 2

11
$\begingroup$

The original reference is, as far as I can tell, this one (following a citation in Kronmal 1993, a very under-read paper imho):

Neyman, J. (1952) Lectures and Conferences on Mathematical Statistics and Probability, 2nd edn, pp. 143-154. Washington DC: US Department of Agriculture.

The stork and baby data is described and analyzed starting on page 143. Despite (or perhaps because) Neyman introduces them with "Once upon a time an inquisitive friend of mine decided to study the question empirically", these data are clearly fictional.

The storks and babies are followed by a railroads example, whose analytical results are apparently real, but whose raw data was reconstructed to show how the same fallacy might have been at work, "Miss Evelyn Fix was kind enough to prepare Table IV indicating what might have been the raw data [...]"

On the other citations: the data from Matthews is from about 50 years after but does seem to have the same structure as Neyman's. It is (I presume) real, and seems to have been collected independently. I cannot find a searchable version of Yule, so despite a personal weakness for hunting through old statistics textbooks, I have not found the time to search it. Perhaps a bird will bring the reference to us.

$\endgroup$
1
  • $\begingroup$ This also agrees with Nick Cox’s comment, since Jerzy Neyman was Polish and this is popular in Polish folklore. $\endgroup$
    – Tim
    Feb 10, 2021 at 17:51
1
$\begingroup$

This (http://www.nieuwarchief.nl/serie5/pdf/naw5-2010-11-2-134.pdf) dutch magazine article mentions

G.E.P. Box, W.G. Hunter en J.S. Hunter (1978),Statistics for Experimenters: An Introduction to Design, Data Analysis, and Model Building, New York: John Wiley, p. 8

as the first example. Box et al. apparently use a data set from Oldenburg, Germany from the thirties (which is also analyzed in the magazine article).

$\endgroup$
2
  • 1
    $\begingroup$ I note that the data is the same place of several years rather than the OPs recollection that it concerned different places at the same time. Also the data are much older than the textbook. (The same is true of the Sies piece, of course). $\endgroup$ Feb 10, 2021 at 16:47
  • $\begingroup$ In the boook by Box, Hunter and Hunter the give the following as the reference for the data in the plot: "Ornithologische Monatsberichte, 44, No. 2, Jahrgang, 1936, Berlin, and 48, No. 1, Jahrgang, 1940, Berlin, and Statistiches, Jahrbuch Deutscher Gemeinden, 27-22, Jahrgang, 1932-1938, Gustas Fischer, Jena". $\endgroup$
    – Phil
    May 12 at 12:09

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.