Here you see my scruffy sketch of Egon drawn 20 years ago for the frontispiece of my book, “Error and the Growth of Experimental Knowledge” (EGEK 1996). The caption is
“I might recall how certain early ideas came into my head as I sat on a gate overlooking an experimental blackcurrant plot… –E.S Pearson, “Statistical Concepts in Their Relation to Reality”.
He is responding to Fisher to “dispel the picture of the Russian technological bogey”. [i]
So, as I said in my last post, just to make a short story long, I’ve recently been scouring around the history and statistical philosophies of Neyman, Pearson and Fisher for purposes of a book soon to be completed, and I discovered a funny little error about this quote. Only maybe 3 or 4 people alive would care, but maybe someone out there knows the real truth.
OK, so I’d been rereading Constance Reid’s great biography of Neyman, and in one place she interviews Egon about the sources of inspiration for their work. Here’s what Egon tells her:
One day at the beginning of April 1926, down ‘in the middle of small samples,’ wandering among apple plots at East Malling, where a cousin was director of the fruit station, he was ‘suddenly smitten,’ as he later expressed it,with a ‘doubt’ about the justification for using Student’s ratio (the t-statistic) to test a normal mean (Quotes are from Pearson in Reid, p. 60).
Soon after, Egon contacted Neyman and their joint work began.
I assumed the meanderings over apple plots was a different time, and that Egon just had a habit of conducting his deepest statistical thinking while overlooking fruit. Yet it shared certain unique features with the revelation when gazing over at the blackcurrant plot, as in my picture, if only in the date and the great importance he accorded it (although I never recall his saying he was “smitten” before). I didn’t think more about it. Then, late one night last week I grabbed a peculiar book off my shelf that contains a smattering of writings by Pearson for a work he never completed: “Student: A Statistical Biography of William Sealy Gosset” (1990, edited and augmented by Plackett and Barnard, Clarendon, Oxford). The very first thing I open up to is a note by Egon Pearson:
I cannot recall now what was the form of the doubt which struck me at East Malling, but it would naturally have arisen when discussing there the interpretation of results derived from small experimental plots. I seem to visualize myself sitting alone on a gate thinking over the basis of ‘small sample’ theory and ‘mathematical statistics Mark II’ [i.e., Fisher]. When nearly thirty years later (JRSS B, 17, 204 1955), I wrote refuting the suggestion of R.A.F. [Fisher] that the Neyman-Pearson approach to testing statistical hypotheses had arisen in industrial acceptance procedures, the plot which the gate was overlooking had through the passage of time become a blackcurrant one! (Pearson 1990 p. 81)
What? This is weird. So that must mean it wasn’t blackcurrants after all, and Egon is mistaken in the caption under the picture I drew 20 years ago. Yet, he doesn’t say here that it was apples either, only that it had “become a blackcurrant” plot in a later retelling. So, not blackcurrant, so, it must have been apple, putting this clue together with what he told Constance Reid. So it appears I can no longer quote that “blackcurrant” statement, at least not without explaining that, in all likelihood, it was really apples. If any statistical sleuths out there can corroborate that it was apples, or knows the correct fruit that Egon was gazing at (and, come to think of it, why couldn’t it have been both?) I’d be very grateful to know [ii]. I will happily cite you. I know this is a bit of minutia–don’t say I didn’t warn you [iii]. By contrast, the Pearson paper replying to Fisher is extremely important (and very short). It’s entitled “Statistical Concepts in Their Relation to Reality”. You can read the paper HERE.
[i] Some of the previous lines, and 6 following words:
There was no question of a difference in point of view having ‘originated’ when Neyman ‘re-interpreted’ Fisher’s early work on tests of significance ‘in terms of that technological and commercial apparatus which is known as an acceptance procedure’. …
Indeed, to dispel the picture of the Russian technological bogey, I might recall how certain early ideas came into my head as I sat on a gate overlooking an experimental blackcurrant plot at the East Malling Research Station!–E.S Pearson, “Statistical Concepts in Their Relation to Reality”
[ii] As Erich Lehmann put it in his EGEK review, Pearson is “the hero of Mayo’s story” because I found in his work, if only in brief discussions, hints, and examples, the key elements for an “inferential” or “evidential” interpretation of Neyman. So I should get the inspirational fruit correct.
[iii] I’m not saying I know the answer isn’t in the book on Student, or someplace else.
Fisher 1955 “Scientific Methods and Scientific Induction” .
Pearson E.S., 1955 “Statistical Methods in Their Relation to Reality“.
Reid, C. 1998, Neyman–From Life. Springer.