ASA Connect

 View Only
Expand all | Collapse all

Gosset & Guinness

  • 1.  Gosset & Guinness

    Posted 01-02-2018 14:50
    Having very recently visited the 6 (or so) story Guinness factory (or Storehouse) in Dublin I was disappointed to find that the displayed extended history of the firm makes no mention of WS Gosset. One would suppose that they would want visitors to know a major scientist had worked at Guinness and that quality control issues for beer provided the motivation and the setting for a major advance in statistics.  The ASA ought to bring this omission to their attention.

    Malcolm Sherman

  • 2.  RE: Gosset & Guinness

    Posted 01-03-2018 03:29
    The IBS had their 2008 conference in Dublin to mark the centenary of the t-test. Some info can be found in an old newsletter. There is apparently a plaque commemorating his work in the storehouse.

    Jake Olivier
    Associate Professor
    University of New South Wales

  • 3.  RE: Gosset & Guinness

    Posted 01-03-2018 04:11
    A statistician a major scientist? When and where? Not exactly Watson and Crick, and I know plenty of PhD scientists in medical/health science, who didnt know who w&c we're before I explained to them.

    Also, Gossett was forbidden by Guinness from publishing, why he used the psuedonym. I don't think Guinness would like to show off how their employee published against company regulations.

  • 4.  RE: Gosset & Guinness

    Posted 01-03-2018 08:22
    ​Interesting, I was there in 2011 and was very excited when I saw the plaque on the wall with his name and contribution.
    Granted it wasn't a huge plaque and it could easily be missed, but it was definitely there, I had my picture taken with it!

    Robyn Lee-Stubbs
    Senior Statistician

  • 5.  RE: Gosset & Guinness

    Posted 01-03-2018 11:50
    I was also there in 2010 and see the attached pictures.  I think you may have missed it.

    Rebecca Pierce
    Ball State University

  • 6.  RE: Gosset & Guinness

    Posted 01-04-2018 05:38
    It is absolutely no surprise to me that the photos of the plaque for Gossett says nothing about him being a scientist or even a statistician given he was forbidden to publish by Guinness. He graduated in both chemistry and mathematics at Oxford before joining Guinness. Statistics was not even a recognised field at the time. As myself both a brewery and a  biochemist turned biostatistician, I'm pretty sure Guinness hired him primarily on his chemistry background. His mathematics background would become a huge asset to the field of statistics, as well as his place of work, very suitable for design of experiments ((DOE).

  • 7.  RE: Gosset & Guinness

    Posted 01-03-2018 11:31
    Since my post yesterday I have received a personal reply from someone
    who saw a Gossett plaque when he visited the Guinness Storehouse in
    2008. While I was consciously looking for something about Gossett, it is
    also possible that the display is still there and I just missed it.

  • 8.  RE: Gosset & Guinness

    Posted 01-03-2018 14:27

    Dear Malcolm, Jacob, and Rebecca,

    Malcolm, by now you've seen the post and photo of the plaque honoring Gosset. The plaque is
    displayed in The Barley Room of the Storehouse, on the back wall.  I was a member of the group at
    IBS 2008 who advocated for the plaque.  Guinness (Diageo), IBC, and members of the Irish
    Statistical Society gathered together with 5 of Gosset's grandchildren and grandnephews
    and nieces for the unveiling.

    (BTW the plaque has at least three errors: Gosset was "Head Brewer", not chief; his famous test is called
    "Student's" t-test, not Student t; and he worked as an experimental brewer, agronomer, and statistician--
    not as a chemist.  I have lobbied for change but not yet with success!)

    Guinness (Diageo) is well aware of their great scientist and pioneering statistician.  At IBS 2008
    there was a special session on the centenary of "Student's" t, with over 300 attendees. Talks were
    given by David Cox, Stephen Senn, James Hanley and me (Stephen Ziliak).  Guinness archivists
    were quite helpful in making that session a reality.

    In 2011, at the European Historical Economics Society meetings, I gave a keynote on "Guinnessometrics"
    in the old Fermentation Room of the Guinness Storehouse, with lots of support from Guinness (Diageo).
    (see the link, below)

    Jacob, I don't understand the sarcasm behind your "not exactly Watson and Crick" comparison.
    (And I note that you didn't mention the highly empirical Rosalind Franklin, without whom W&C had little:
    see the excellent bio by Brenda Maddox, for example, detailing W&C's exploitation of Franklin's findings
    and gender.)

    I have been working in the Gosset archives since 2005, and find a quite different story to be true.

    Yes, Gosset was a major scientist, a major scientist who happened to invent or inspire half of modern
    statistics during the course of his work. I have written extensively about Gosset and "Guinnessometrics" in
    a number of articles and in my book (with Deirdre McCloskey), The Cult of Statistical Significance
    Most of the second hand histories after Fisher have been wrong about Gosset. My work attempts to
    correct those errors, with implications for today's practice (design of experiments, statistical significance
    testing, replication, reproduction, economic approach to real error estimation, etc).  Your pint of Guinness
    might feel like magic but that's in reality the magic of great science.

    Many people have been led to believe the rumor that Gosset and other Guinness researchers were not
    allowed to publish. That is plain false, as I show in my work.  In truth, Gosset and colleagues were sent
    occasionally on sabbatical to up their research game, and Guinness allowed Gosset and others to work
    on publications during (rare) slow hours at the brewery.  Gosset published 21 articles - 14 in Biometrika -
    and every one of them was distributed to the Guinness Board and Managing Director. 

    Why was he called "Student"? The Guinness rule was: don't mention Guinness, don't mention beer, and don't
    mention your own surname.  Secrecy was standard practice for British and Irish companies protecting proprietary rights.
    (Other Guinness researchers appearing in Biometrika include "Sophister" and "Mathetes".)  We don't know
    for sure why Gosset chose "Student" for a pen name but as I note in my book, the first notebook he used while
    on sabbatical at UCL-Biometrika Lab in 1906 was called "The Student's Science Notebook" (Ziliak and McCloskey
    2008, p. 213).

    Here are some links of possible interest:



    Stephen Ziliak

    Professor of Economics
    Roosevelt University (Chicago)

    Conjoint Professor of Business and Law
    University of Newcastle (Australia)

  • 9.  RE: Gosset & Guinness

    Posted 01-04-2018 05:54
    Hi Stephen,

    I know who you are - i am a strong critic of statistical significance myself.

    I'm passionate about the history of science. I've had lunch with James Watson when I worked at harvard (he was visiting from cold spring s harbor - I'm not that old) he reluctantly signed my copy of the double helix. I've also read the books of Francis Crick, Maurice Wilkins, and also Brenda maddox's the dark lady of DNA. None of that is at all relevant to what I was talking about. What I'm talking about is that the wider society doesn't really acknowledge statistics as useful as it actually it to science. That is statisticians do not receive the recognition they deserve so it would be no surprise at all that Guinness does not acknowledge the contributions of Gossett to society, to brewing science, or science generally.

  • 10.  RE: Gosset & Guinness

    Posted 01-06-2018 12:28

    Hi Jacob,  thank you for clarifying! Glad to hear it. I share your frustration with the all-too-common tendency of many scientists and critics to speak of statisticians as if our ken are not scientific. It's a simple historical error, among other things.  Statistical theory and methods have developed precisely because the developing scientists have found a need for them. If Student's statistical methods are unscientific, then so too are telescopes and satellites and post-Copernican astronomy unscientific. 

    By the way, Pearson pere's first name was Karl, not Charlie.  And Gosset's surname has only one "t", as in one t-distribution!

    My favorite way to put it is this: 

    "Gosset has one t, Hotelling has two." (Hotelling's T-squared)! 



    Stephen Ziliak
    Prof of Economics
    Roosevelt University

  • 11.  RE: Gosset & Guinness

    Posted 01-04-2018 06:12
    "The Guinness rule was: don't mention Guinness, don't mention beer, and don't 
    mention your own surname. Secrecy was standard practice for British and Irish companies protecting proprietary rights'

    It is no secret that Gossett went on sabbatical to work with Charlie Pearson. Why he published under Student? As you said, he couldn't use his own name or Guinness. With what we know, this isn't a mystery.

  • 12.  RE: Gosset & Guinness

    Posted 01-05-2018 12:18
    I have been curiously reading the blogs about Gosset (William Sealy) and found myself in a spot to respond.  First, Jacob Humphries mentioned in one of his blogs that Gosset took a sabbatical to work with Charlie Pearson.  I am sure Jacob did not mean Charlie, he meant the great Karl Pearson under whom Gosset did a two-year sabbatical research.  The only two famous statisticians of yester years with the last name Pearson were Karl and his son E.S. 

    One interesting anecdote about Gosset is that he was a friend of both Karl Pearson's and R. A. Fisher's, which was some difficult feat given the fact that these two giants only had name calling relations.  In any case, Gosset's work on his t-distribution, design of experiments, and quality control had important impact on Fisher, Taguchi Genichi, and Madhav Phadke's (Bell Labs.) works on quality control and design of experiments. 

    Gosset left the old Guinness in Dublin at the age of 59 and joined the new Guinness in London's Park Royal location as the Head Brewer in charge of quality control and design of better experiments.  He died two years later at the age of 61.

    There are many myths and stories about why Gosset took the name Student to propose his t-test as Student's t-test.  It is true that Guinness had restrictions on names in publications by their scientists because earlier one scientist presumably disclosed some stuff about their program; but they actually tried to encourage them to give talks and publish on their research without disclosing their real names.  Remember, Gossett published about a dozen and a half articles in Karl Pearson's journal Biometrika.  Over forty years ago I met H.O. Hartley in a gathering and many, including myself, tried to encourage that small-in-size-but-giant in intellect statistician who fondly knew them all.  I asked whether he had any theory why Gosset published under the pseudonym Student.  He chuckled and said according to Gossett, his t-test was so simple that even a student should be able to do it without much struggle.  Is that the reason behind the name Student's t-test? Quite likely! 

    Ajit K. Thakur, Ph.D.
    Retired Statistician

  • 13.  RE: Gosset & Guinness

    Posted 01-06-2018 12:36

    Thank you for these comments, Ajit.  My claims are based on archival research from at least a dozen libraries and same number of years of research on Gosset, Guinnessometrics, and the development of modern stats.  You might be interested in my book and articles posted here:

    The Cult of Statistical Significance

    Stephen Ziliak
    Prof of Economics
    Roosevelt University

  • 14.  RE: Gosset & Guinness

    Posted 01-10-2018 15:11
    "I am sure Jacob did not mean Charlie, he meant the great Karl Pearson under whom Gosset did a two-year sabbatical research."

    I intentionally wrote 'Charlie'. K Pearson was British, and Charlie, or Charles, is an anglicised version of Karl. Years ago, when I was a psychology undergrad, I did a semester in Australia and found psych professors there referring to Carl Jung as Charlie Jung. I am pretty sure I read somewhere that at some point Pearson changed his name to Karl.

    I just did a quick google search and found pages saying his name was Carl but when he enrolled at University of Hiedelberg they misspelled his name Karl at that spelling stuck.

    Jake Humphries

  • 15.  RE: Gosset & Guinness

    Posted 01-03-2018 14:42
      |   view attached

    Gosset is featured in a wall plaque very near the start of the tour one has to pay for. It does not say how influential he was (if one does not already know).  

    I have visited that brewery several times. Once I made my way to their in-house library of records and asked to see information on/from Gosset (like letters to/from Fisher). This request drew a blank from their reference librarian. She never heard of Gosset. She looked but found nothing in their records about him. I was disappointed.


    DSCN0968kenGosset.pdf   124 KB 1 version