The Oxford Dictionary of Statistical Terms
The first edition of this dictionary appeared in 1957 and contained 1,700 terms. It became a well-respected reference book. The revision for the current sixth edition started in 1998. A website was constructed where all the terms of the fifth edition were listed. The statistical community suggested 320 terms for elimination and 1,067 terms for addition. It was decided to eliminate 265 entries and to add 640 new entries. The sixth edition now contains 3,540 terms. For some entries a reference to the literature is given. The list of references is quite long (pp. 439-498) and it is one of the positive features of the dictionary. The readers and users of the dictionary are invited to send their comments and suggestions to a website of the International Statistical Institute. The address of the website is introduced in the preface.
I would just like to make the comment that I am puzzled by the explanation of the term Rayleigh distribution presented on page 339: “A χ2 distribution with two degrees of freedom, so called because it was considered by Rayleigh in some physical situations”. In statistical textbooks and papers we read something different. For example, in Lindgren B. W. (1993): Statistical Theory, 4th edition, Chapman and Hall, p. 186, we find the following information: “if X and Y are iid N(0,σ2) random variables, then R=√(X2+Y2) has the Rayleigh distribution. Since X2+Y2 has the χ2distribution with two degrees of freedom, the distribution of R cannot be χ¬¬¬¬2 ”. Despite this small point, I must say that the book covers a broad area of mathematical statistics and it is a useful reference for fast orientation on statistical terms.