This book review summarizes shortcomings in a statistics textbook for experimental science students. The reviewer notes that while the book provides good foundational content, it fails to adequately address more advanced statistical techniques needed to isolate variables, test differences, and predict results. Specific technical errors are also identified, such as mistakes in the correlation treatment and improperly distinguishing statistics from parameters. The reviewer concludes the book misses an opportunity to convey to students that both science and statistics continue advancing.
This book review summarizes shortcomings in a statistics textbook for experimental science students. The reviewer notes that while the book provides good foundational content, it fails to adequately address more advanced statistical techniques needed to isolate variables, test differences, and predict results. Specific technical errors are also identified, such as mistakes in the correlation treatment and improperly distinguishing statistics from parameters. The reviewer concludes the book misses an opportunity to convey to students that both science and statistics continue advancing.
This book review summarizes shortcomings in a statistics textbook for experimental science students. The reviewer notes that while the book provides good foundational content, it fails to adequately address more advanced statistical techniques needed to isolate variables, test differences, and predict results. Specific technical errors are also identified, such as mistakes in the correlation treatment and improperly distinguishing statistics from parameters. The reviewer concludes the book misses an opportunity to convey to students that both science and statistics continue advancing.
On page 59 it is stated that since table when F is defined to be greater than
BOOK REVIEWS the book is for students in experimental one is not correot. The discussion of the science, in which "controlled" conditions handling of data suspected of being had is its. However, with a few exceptions to obtain, little space will be devoted to the entirely out of date. Chauvenet's re- be noted, this is a good hook for those statistical treatments of extraneous fac- jection criterion, the only one discussed, whose work consists of making repeat tors. "If and mrhen the student needs rejects data a t the 50 per cent level of measurements and studying their dis- them, he e m find these treatments in the significance for large sample sizes, and will tributions. It is not a good book for literature" (itslics supplied). With this never reject any measurements, no matter anyone who wishes to go further and st,roke of the m n the author dis~osesof how had they are, for sarnplesizesless than isolate sources of variation, estimate the five! The author claims that the chi importance of extraneous variables, test square test requires mare than about five the significance of the difference between thelast thirty years! degrees of freedom to be satisfactory; two or several means, or predict a. result At times it is quite evident that the yet he freely uses the normal approxima- based on a.least-squares line. author is writing out of his field. The tion to the binomial distribution, and this Students m.ho use this text uill be led treatment of correlation contains serious is algebraically identical to a chi square to expect much less of statistics than they mistakes. The interpretation of the F test with one degree of freedom. There is no need to limit the degrees of freedom of chi square. Throughout the book the author has been careleas about distinguishing between a statistic and a parameter, as on page 124 where both Student's t statistic and the F-ratio statistic are called parameters, See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
and on page 179 where the same symbol
is used for both the theoretical and the observed binomial probabilities. But per- haps most serious of all is the way statisti- cal design of experiments is undercut by Downloaded via 122.161.241.209 on October 11, 2022 at 12:22:05 (UTC).
being described as referring only to how
large the experiment must be to give s specified precision. Science, ss is pointed out several times in the book, is "open-ended," i.e., subject to further improvement. So is the applicstion of statiisties of experimental science. Professor Parratt has missed an opportunity to tell this to students. LLOYD5. NELSON General Electric Lamp Division Cleveland. Ohio
Introduction lo Chemical Engineering
L. Bruce Andwsa, University of Nebraska, Lincoln, and L e a a d A. Wend, Lehigh University, Bethlehem, Pennsylvania. McGraw-Hill Book Co., Inc., New York, 1961. xii + 364 pp. Figs. and tahles. 16 X 23.5 cm. $9.50. This book, designed to serve as an introduction to chemical engineering, enters a field which is crowded. How- ever, though chemical engineering texts are numerous, few are designed aa intrc- durtions, the purpose outlined by the authors of this volume. The choice of text material is rather interesting. A brief description of the functions of a chemical engineer is followed by s. section of useful mathematical methods including a discussion of digital computers. The presentation and corre- lation of data are items which are not normally given and the authors are to be commended for doing this. Enough physical and chemical prin- ciples are given to serve the subsequent sections on material and energy balances. The section on balances is the best part of the book. The last part is devoted to procees industries of the inorganic and organic type. Extensive use is made of problems thus contributing to the quanti- tative aspect of chemical engineering. (Continued on page AS16)
A314 / Journal o f Chemicol Education
BOOK REVIEWS
I t is doubtful wherc a course, using this
t,ext ns a basis, ran be fitted into an already overcrowded mmirulum. The scope of the book is not m r h that it may he suh- stituted in t,he customary stoirhiometric course. Hence, it is recommended as a, supple mentary text, not as a substitute for the excellent lmoks in the stoichiometric field.
Selected Studies in Chemical Kinetics:
Thirty-Fifth Annual Priestley Lectures Farrington Daniels, University of Wis- consin, Madison. Published by Phi Lambda Upsilon and the Department of Chemistry, Pennsylvania State Uni- versity, University Park, 1961. vi + 123 pp. Figs. 21.5 X 27.5 cm. Paper bound. $2.75.
The Priestley lecturers are more than
recognized authorities; they are m m known t o he able to make their subject understood by undergraduates. Professor Daniels has been well chosen on both counts. As can be expected, this is mainly a N-0 story, hut a story n.ell told, excellent as supplemental reading for m y physical chemistry course. W.F.K.
Toxicology: Mechanisms and
Analytical Methods. Volume 2 Edited by C. P. Slmoa~t,University of Edinburgh, Scotland, and A. Stolnmn, Connecticut State Department of Health, Hartford. Academic Press, Inc., New York, 1961. rvi + 921 pp. Figs. and tablea. 16.5 X 23.5 cm. $2.5. T h i s volume is sn extension of the Volume 1 of this series produced by a team who are experts in their field. I n the second volume, the authors have continued the same format as Volume 1. The book contains adequate descriptions of general procedures of analyses, along with pertinent bihliography so that an analyst, somewhat versed in technique, may satisfactorily ohtain the desired results. The section which relates to enzymatic activity is stimulating. Both t,he author and the reviewers encourage pursuit of such studies, since they should be valuahk in understanding the basic mechanism of toxicological activity. However, a r o r d of caution is offered that such chanees
this respect over a proper period of time.
The reviewers were impressed wit,h (Continued on page .4318)