Talk:William Sealy Gosset
|This article is of interest to the following WikiProjects:|
states that the reason William Sealy Gosset used a pseudonym was not that they forbid publications but that they did not want competitors to know they got an advantage through hiring a statistician.
- The article already says that. Michael Hardy 18:11, 31 August 2006 (UTC)
- Oh... I see; it's just a bit different. Michael Hardy 18:12, 31 August 2006 (UTC)
Is the formula for z correct?
"Gosset's statistic was z = t/√(n - 1)."
Given that Gosset's statistic was z = x-bar/SD, and given that the standard t formula is t = x-bar/SD√(n - 1), shouldn't the formula in the essay be z = t*√(n - 1)? Or, put another way, shouldn't the formula be t = z/√(n - 1)?
NiallB 01:18, 31 July 2007 (UTC)Niall Bolger
- [from Terry0051] Just a note for the record (while in checking mode...) that this (now somewhat old) talk-page suggestion did not seem to be correct, and that the article text here was already correct. "Student" described z (at p.24 in the reprint of his 1908 paper) as "the value of the mean, measured from the mean of the population, in terms of the standard deviation of the sample". Both this, and the comparison between "Student"'s table and newer tables that results from it, match ok with the formula z = t/√(n - 1) as given. Terry0051 (talk) 20:33, 16 April 2009 (UTC)
Relationship of Golsset to K. A. Pearson, and R. A. Fisher.
The present text refers to the giant egos of Pearson & Fisher,which I do not dispute :) But it wants a reference.
That Gosset was able to be on good terms with each of them was an amazing accomplishment.
The reference to Gosset's accomplishment with this, and some details of Pearson & Fisher, can be found in
Box, Joan Fisher, _R. A. Fisher, The Life of a Scientist_, 1978. Sorry, I don't have a page ref.
but she spends a good deal of time on the p;personalities involved. Gosset was very helpful to Fisher inasmuch as Gosset provided an informal access to Pearson as well.
Also in this book, J. F. Box describes the reason for the pseudonym completely differently than others. I rather prefer the cause as Guinness' concerns about competitive advantages, because it fits the other facts of the situation much more closely.
Academic relation (if any) with Airy dubious if not impossible
The stated relation with G B Airy (said to be Gosset's 'academic advisor') surely needs a citation, and seems very dubious if not impossible in view of their ages. Airy's dates were 27 July 1801 to 2 January 1892, Gosset was born June 13 1876, and Airy died at age 91 when Gosset was still under 16 years of age. The stated 'fact' is clearly dubious. Terry0051 (talk) 10:13, 15 April 2009 (UTC)
- Hmmm, well spotted. I suspect the source was the MacTutor biography of Gosset, usually pretty reliable but here fairly clearly in error as it claims Gosset studied under Airy while at New College Oxford, but Airy was in fact at Trinity College Cambridge. I suspect this page from The Cult of Statistical Significance may have some bearing on the confusion:
"I had learnt what I knew about the errors of observations from Airy" (i.e. from the astronomer G.B. Airy's Theory of Errors of Observations (1861).
- I've therefore removed the 'academic advisor' entry from the infobox. Qwfp (talk) 20:01, 15 April 2009 (UTC)
What a cool dude!
So why did he choose the pseudonym 'Student'?
- I'm not sure anyone really knows, but:
Gosset adopted the pseudonym ‘Student’ when he published his first paper in 1907 (we may now guess) precisely because he saw himself at that time as a student of statistics.
|— Peter Petocz & Eric Sowey, Statistical Diversions. Teaching Statistics 2006. doi:10.1111/j.1467-9639.2006.00246.x|
I thought there was an image of Gosset's signature on the page, though I can't find a record of its deletion. I normally wouldn't care, though his was one of the best signatures I have ever seen!Ninahexan (talk) 06:24, 15 February 2011 (UTC)
I just edited out a sentence in the article claiming (without giving a source) that the t-test was first invented by R.A. Fisher. The t-test was published in 1908 by Gosset. R.A. Fisher was only 18 at that time, so this makes it highly unlikely that he could have invented t-test while at secondary school. In fact David Salsburg in The lady tasting tea: how statistics revolutionized science in the twentieth century (p. 31) gives the explanation behind this misconception (that Fisher invented he t-test before Gosset). Fisher actually re-invented the t-test while an undergraduate at Cambridge in 1912. His tutor quickly pointed out to Gosset's 1908 paper which Fisher was unaware of and then proceeded to introduce Fisher to Gosset (who was 14 years his senior) and the two became friends. 18.104.22.168 (talk) 11:41, 4 May 2012 (UTC)