|
Essentials of Biostatistics
|
|---|
Second edition. Robert C. Elston and William D. Johnson. 328 pages. Philadelphia: F. A. Davis; 1994. $31.95.
Many textbooks on biostatistics at the elementary, intermediate, and advanced levels are currently available. The present book is an ordinary member of this group; it makes no extraordinary contributions. It addresses the standard sequence of topics with many educational, interesting, and applied problems that are capable of generating intellectual curiosity in the fertile minds of serious students of biostatistical science. Elementary and intermediate topics are lucidly discussed and illustrated. However, advanced topics are treated superficially and without proper caution; thus, they are likely to provide misleading, biased, and nonrobust conclusions. The book presents materials at various levels, which may be a detriment for beginners because proper applications of advanced specialized techniques need considerable theoretical and applied statistical sophistication.
Specifically, chapter 11, which deals with some specialized advanced techniques, fails to reflect proper scientific wisdom. Statistical inferences derived from the applications of these techniques are often highly controversial and often provide biased inferences. It is dangerous to introduce specialized advanced techniques in an elementary textbook without providing proper guidance.
Meta-analysis, which is covered in chapter 12, has become very troublesome. It requires a much more meaningful discussion than the one this book provides. Related to meta-analysis is the combination of P values, and treatment of this topic is unsatisfactory. Limited space prevents me from covering specific critical points pertinent to this technique in this review.
Modern statistical science has been enjoying accelerated advances in recent years because of the ever-increasing power of computer technology. This technology permits difficult ideas of theoretical statistical science to be elegantly illustrated in an intellectually exciting, creatively challenging, and experientially rewarding way. Sadly, this book fails to reflect the fruits of these advances. It is a traditional book reflecting traditional statistical theories, methods, applications, and practices in a pedantic style.
On the whole, this book may be considered satisfactory for certain beginning students of statistical science; the advantages outweigh the disadvantages. The real strength of this book is in providing problem sections and further reading references. I strongly believe that statistical science deserves more than what most books in the field have offered. The power of computer technology has totally changed the destiny of statistical science; most books do not reflect this.
Writing a textbook on the difficult subject of biostatistical science is extremely challenging. Consequently, the authors deserve great compliments for publishing a book that contains an enormous amount of useful information. This text should serve as a good addition to the personal libraries of statistical instructors as well as consultants.
Article
