Etymology of "t" in t-test and t-distribution

Question

I am aware of the fact that Gosset came up with t-distribution, but what is the etymology of "t"? How did "t" end up in t-test and t-distribution?

score 10 · Answer 1 · answered Jan 15 '12 at 18:42

10

It appears to have been done by Fisher as he was developing a more general theory of testing, although perhaps by Student in a letter to Fisher a couple of years before. Originally Student denoted it by $z$, although it wasn't actually in its current form at that time (that had to wait for Fisher.) If you have access, The American Statistician (1979) has a paper on this. If you don't, some brief comments are in http://jeff560.tripod.com/s.html.

answered Jan 15 '12 at 18:42

jbowman

31,550
8
54
107

Could you edit in a brief summary of the main points of the linked sites into your answer, in case there's issues with them later? (For example, for some reason I cannot successfully connect to the tripod.com site at the moment.) – R.M. Apr 14 '17 at 14:15

score 6 · Answer 2 · edited Apr 13 '17 at 12:44

6

According to H.A. David's paper linked here, "Student's t" and "t" were first used in R.A. Fisher's 1924 paper On a Distribution Yielding the Error Functions of Several Well Known Statistics. The first occurrence is on page 809 where we see Fisher refer to "Student's curves" and uses the variable t.

edited Apr 13 '17 at 12:44

Community

1

answered Jan 15 '12 at 18:34

score 1 · Answer 3 · answered Jan 15 '12 at 18:18

1

I believe it's because "t" is the last letter of both Gossett and Student, which was his pen name. Why choose the last letter? Search me.

answered Jan 15 '12 at 18:18

rolando2

11,645
1
39
60

2

Gossett is a very common typo, but he really was Gosset. https://global.oup.com/academic/product/student-9780198522270 – Nick Cox Jan 29 '16 at 11:48

Etymology of "t" in t-test and t-distribution

3 Answers3