Download View/Open

First Course in Probability and Statistics. By J. NEYMAN. Henry Holt and Company, New York.
350 pages, illus. 1950.
•
guage and notation of the mature mathematical
reaatr tio
ol t reacquaintance awarilthythe t P
a t usrs coam
Nwehiylo hnaans
lNYONE
NsY
tia etOsi ssNot sw
statistician; the beginner need not let this deter
him. The vocabulary and symbolism must be
naturally
statistics
learned sometime by a prospective statistician and
ment of this subject, combined with a style of
it may as well be at the start of his career. The
writing that makes it come very much to life on
exposition is clear and is accompanied by many
the printed pages. He will not be disappointed.
examples and exercises. Biologists should be deMoreover, this book is unique among its fellows in
lighted
to learn that a chapter of 67 pages is dethat statistics is presented as a branch of Probavoted
to
probabilistic problems of genetics, and
bility Theory.
that
other
applications to biological statistics apAccording to the preface, it is intended as a text
pear
throughout.
for " (i) students who would like to take just one
As is to be expected from the title, the book is
course in mathematical statistics for purposes of •
primarily
an exposition of probability theory, with
general education, (ii) prospective future matheprobability
defined as a relative frequency, tomatical statisticians and (iii) students who spegether
with
applications
of that theory to the gencialize or intend to specialize in one of the fields
eral
problem
of
testing
hypotheses
and some typiof application and need mathematical statistics as
cal
problems
encountered
in
laboratory
experimena useful tool in their own studies." He visualizes
tation
and
in
sampling.
The
concepts
underlying
the first two categories as being freshmen or sophodecision functions play a prominent part and the
mores, and the third as consisting of senior or
power of a test is given adequate attention. Congraduate students who have more extensive mathesiderable
material ordinarily given in texts on
matical training and a more mature viewpoint.
mathematical statistics is not even mentioned, conIt seems to this reviewer that those in the first
sequently there is here no threat to the demand
category would find the book rather heavy going,
for such treatises. Most statistical taxonomists will
but if they will expend the necessary effort, they
doubtless classify it as a book on probability rather
ill be rewarded by an insight into the fundathan on statistics per se. At the same time, it conentals of statistical thinking that can hardly be
tains much that ought to be included in texts on
obtained from any other current work on elemenmathematical statistics but seldom is.
tary mathematical statistics. But, human nature
This reviewer recommends the book highly to
being as it is, many will probably continue to prereaders of this journal who have the wish and the
fer a more superficial survey of a larger variety of
fortitude to get at the heart of the subject. It is
statistical techniques to a searching examination of
not easy fare ; anyone seeking a pablum, one that
the foundations upon which those techniques rest.
will impose no strain on his digestive processes will
Professor Neyman's presentation of probability
not find it here. But, for that matter, he will not
theory, although restricted to the viewpoint of
find it anywhere else either.
probability "as a mathematical model of relative
frequencies, observable in a long series of trials,"
Walter A. Hendricks
is excellent. He does not hesitate to use the lan-
A
w
37