April 15, 1922] 



NATURE 



473 



Ogether the substance of important controversial 

 ipers on this subject published from time to time by 

 >f. Godfrey Thomson during the war. There are, 

 >wever, several other new insertions. There is a long 

 id lucid chapter on the elementary theory of proba- 

 Ijlity and a shorter and more abstruse one upon skew- 

 ss and heterogeneity in psychophysical data — a 

 ipter in which, as indeed all through the volume, 

 rof. Karl Pearson's views and manner of approach 



plainly reflected. 

 In the present stage of its development the book is 

 interesting though composite production. From 

 title one would expect some description of the 

 mt methods of individual measurement — methods 

 )r assessing intelligence and for estimating the capa- 

 ties and attainments of the mind. Instead we have 

 itical prolegomena upon statistical adjuncts, and the 

 iters have in view, not so much general mental 

 |ting, as the testing of the mental tests themselves. 

 ie book opens with a general chapter headed " Mental 

 isurement," which seems to promise the broad 

 Turvey characteristic of a general text-book, but most 

 of the pages in the second portion of the book, and 

 most of the paragraphs in its two independent prefaces, 

 are simply brilliant contributions to a special con- 

 troversy. The three doctrines chiefly attacked are 

 those connected with the name of Prof. Spearman. 

 It is, however, highly satisfactory to learn that Dr. 

 Brown finds elements of truth in the views that he 

 disputes, and feels himself to be " more convinced than 

 ever that the work of Prof. Spearman's correlational 

 psychology is epoch-making in its significance." 

 The heterogeneous character of the volume has been 

 verely criticised by at least one eminent statistician.^ 

 Uut perhaps the best defence to this criticism is the 

 simple circumstance that at present the whole subject 

 of mental measurement has itself arrived at a some- 

 what heterogeneous stage. In any case, the mixed 

 ind controversial quahty of the book does not lighten or 

 Ueviate the difficulties of a topic already intrinsically 

 perplexing. The treatment is of necessity technical. 

 From first to last the pages of the book are dotted with 

 algebraic formulae and symbols, and the description of 

 methods is at times extremely condensed. Indeed, 

 to any but a mathematician little but the general gist 

 of discussion can at times be comprehensible. How- 

 \er, a third section to the book is promised, summaris- 

 ing in non-mathematical language both theories and 

 results for the general reader. 



Meanwhile no research student who thinks of em- 

 ploying the statistical methods here described and 

 discussed can ignore this treatise. It is, indeed, con- 



\ G. U. Yule, Brit. Joum. Psych., vol. 12, pp. 100-107 — an article which 

 - Itself of much importance as a contribution to the points at issue. 



x\0. 2737, VOL. 109] 



sidering the intricacy of the subject, a work of great 

 lucidity and compression, and, whether the criticisms 

 urged against previous workers are insuperable or not, 

 objections to their views certainly required statement 

 at length and in detail. 



(2) The volume by Prof. Wilson and Prof. Hoke is 

 altogether different from that by Dr. Brown and 

 Prof. Thomson. It is written for American teachers, 

 and describes in simple phraseology some of the methods 

 in vogue for individual measurement. The chapters 

 consist mainly of an account of the various tests 

 recently standardised for measuring attainments in 

 the chief subjects of the school curriculum — reading, 

 writing, drawing, arithmetic, and the like. A few 

 pages are added on the measurement of intelligence 

 and on -statistical methods and terms. The book is 

 published in the hope of encouraging the teachers 

 themselves to apply to their classes the diagnostic 

 methods that hitherto, for the most part, have been 

 handled by psychologists alone. It is one of the many 

 popular volumes that ha\'e appeared and are likely to 

 appear upon the practical educational applications of 

 these psychological methods. 



Statistical Method. 



A First Course in Statistics. By D. C. Jones. (Bell's 

 Mathematical Series.) Pp. ix-f-286. (London: G. 

 Bell and Sons, Ltd., 1921.) 155. net. 



THE needs of the student of social statistics form 

 the prime consideration in this " First Course," 

 but, as the author states, illustrations have been 

 drawn from all sources, and it will serve very well as 

 a brief introduction for students in other branches of 

 science. The volume has been divided into two parts. 

 Part I. is elementary in character, and in the main 

 can be followed by a reader with little mathematical 

 knowledge. The notions of measurement and of 

 variables are explained, and the conceptions of the 

 frequency distribution, of classification and tabulation 

 are briefly discussed, and a couple of chapters follow 

 on the simpler forms of average and the weighted mean. 

 Dispersion comes next, accompanied by a more detailed 

 discussion of the frequency distribution. The following 

 chapter is on graphs, an unusual feature in this chapter 

 being the inclusion of sections on interpolation and on 

 supply and demand curves. A treatment of the 

 correlation of two variables on simple lines concludes 

 the first part of the book. Part II., though it begins 

 simply, is of a more advanced mathematical character. 

 The first few chapters are on probability, sampling, and 

 probable errors. Prof. Pearson's generalised proba- 

 bility curves are then dealt with, and the method of 

 moments ; two chapters on the normal curve and the 



