
The
University of Michigan mathematician Harry C. Carver founded the
Annals of Mathematical Statistics in 1930, loosely under the aegis
of the American Statistical Association (ASA), with modest financial
support from the ASA. The preface to the first issue was written
by the SecretaryTreasurer and future President of the ASA, Professor
Willford I. King of New York University. In that preface, King boldly
claimed that the ASA had been in the vanguard for 91 years, and in
order to remain there, they needed to include the increasingly complex
mathematical techniques that were then being introduced. Willford
King stated, "For some time past, however, it has been evident
that the membership of our organization is tending to become divided
into two groups  those familiar with advanced mathematics, and
those who have not devoted themselves to this field. The mathematicians
are, of course, interested in articles of a type which are not intelligible
to the nonmathematical readers of our Journal." (King, 1930)
King predicted that the Annals would help serve both groups, and
he expected it to include both theory and applications.
Those
early Annals appear today a bit quaint, filled for the first
few years mostly with review articles (all with handwritten
formulas), unending pages of formulas for moments and semiinvariants
of various statistics, and a few reprinted articles from
other sources. The original articles that did appear were
a curious mix. Articles that we recognize today as of great
historical significance, like Harold Hotelling's 1931 Annals
paper "The Generalization of Student's Ratio" were
exceedingly rare  in fact, I have just named them all.
More typical was a cute little 1933 simulation study by Selby
Robinson, "An Experiment Regarding the ChiSquare Test".
Robinson's simulation (based on coin tosses) verified for
a simple example that Ronald Fisher had indeed been correct
in his correction of Karl Pearson regarding the degrees of
freedom when parameters are estimated.
In
1933 the ASA came under the same overriding concerns for
budget that have recurrently plagued it ever since, and in
December of that year the same Willford I. King (who you
will recall had endorsed the Annals in 1930) led the move
to strip the Annals of its meager ASA subsidy. King had done
an early form of a spreadsheet analysis and claimed that
half the cost of producing the Annals was being subsidized
by nonAnnals subscribers  as he put it, "members,
most of whom are not specialists in mathematics, and hence
find the articles in the Annals not particularly adapted
to their needs." (Hunter, 1996) In fact, King's bookkeeping
was faulty  his budget included a salary for the Annals
Editor when none was being paid (nor, as a matter of principle,
has a salary ever been paid to an IMS editor), and he assumed
there would be no loss of membership in ASA with the demise
of the Annals. But the hero of the day was Editor Harry C.
Carver, who in January 1934 took over the Annals at his own
expense and maintained it without institutional base or support.
By
October of 1934, Carver had evolved an idea for an association
of mathematical statisticians as a base for the Annals, and
despite his earlier experiences he approached the ASA again,
to see if he could arrange for such an association within
the ASA. The ASA was interested, but in the end the interest
was insufficient. On the one hand, the ASA did not want a
new organization dedicated to statistics to start without
their involvement, nor, on the other hand, in the words of
their President Frederick C. Mills, an economist at Columbia
University, did they want to encourage the establishment
within the ASA of "a movement which [might] tend towards
the disintegration of the Association." (Hunter, 1996)
Given his past experience with ASA, Carver was reluctant
to pursue an affiliation further, and he and a number of
likeminded mathematical statisticians, particularly the
University of Iowa's H. L. Rietz, moved forward on their
own. The IMS was officially organized at a meeting at Ann
Arbor on September 12, 1935, with H. L. Rietz as President,
Walter Shewhart as vicepresident, Allen T. Craig as Secretary/Treasurer,
and the three original voting Fellows  a sort of membership
committee  being Burton H. Camp, Arthur R. Crathorne, and
Harold Hotelling. They designated the Annals as the official
journal of the Institute. Later, in 1938 the IMS took over
full financial responsibility for the Annals from Carver.
In
1938 Wilks succeeded Carver as Editor of the Annals and appointed
a stellar editorial board, consisting of Fisher, Neyman,
Cramer, Hotelling, Egon Pearson, Darmois, Craig, Deming,
von Mises, Rietz, and Shewhart. Sam Wilks edited the Annals
for a dozen years, and he transformed the Annals into the
most influential statistics journal in the world. [Extracted
from Stephen Stigler (1996)]
References
Craig,
Cecil C. (1978). Harry C. Carver, 18901977. Annals of
Statistics 6: 14.
Hotelling,
Harold (1931). The Generalization of Student's Ratio. The
Annals of Mathematical Statistics 2: 360378.
Hunter,
Patti W. (1996). Drawing the Boundaries: Mathematical Statistics
in 20thCentury America. Historia Mathematica 23: 730.
King,
Willford I. (1930). The Annals of Mathematical Statistics.
The Annals of Mathematical Statistics 1: 12.
Robinson,
Selby (1933). An Experiment Regarding the Chisquare Test.
The Annals of Mathematical Statistics 4: 285287.
Stigler,
Stephen M. (1996). The History of Statistics in 1933. Statistical
Science 11:244252
