Primer of Biostatistics, 5th ed
Stanton A. Glantz, PhD. 489 pp.
New York, NY: McGraw-Hill; 2001.
US rate: $34.95.
Do we really need another biostatistics textbook? They seem to be being produced hand over fist, and on receiving this one to review, my bias usurped my initial investigation. As a biostatistician, I did not expect to learn anything, and I was especially not looking forward to the same presentation of the same material that purported to do so in a new, different, innovative, applied, or practical manner. Nonetheless, taking the advice of Herbert Spencer not to allow “contempt prior to investigation” to perpetuate my ignorance, I reviewed the book and quickly realized that I was in for a pleasant surprise.
The presentation is really different. Dr Glantz’s target audience is not biostatisticians. There are thorough explanations using clinical terminology without the extensive mathematical proofs that a text for biostatisticians would need. The preface describes the tone as a “confrontational style,” which works well for a reference book for physicians in medical research. Furthermore, although the methods are taught with enough thoroughness to allow the reader to actually do statistical analyses, the language and examples provide the option for the reader to easily learn just enough to become a good consumer of biostatistics. Dr Glantz has met a need that I have known to exist in the medical research community, ie, for a course to allow clinicians to better understand statistical presentations in the medical literature, enable readers to question conclusions on the basis of inappropriate analyses, and improve their ability to work with statistical consultants.
I was completely won over by Dr Glantz’s statement, “Most of the errors (at least as they relate to statistical inference) center on misuse of the t test.” Then, with courage on top of courage, the ANOVA is presented before the t test in the text. This seemingly simple change in order is so important to teaching researchers how to prevent type 1 errors, which is the whole purpose of quantifying deductive inference with statistical inference. Not to mention that this is the true historical chronology of the development of these mathematical concepts.
Although the structure, with questions and exercises at the end of each section, allows it to easily serve as a course textbook, the conversational writing style and organization make the book useful as a reference for a lifetime. It is the best statistics text for nonstatisticians that I have read. Biostatisticians who spend their days consulting and supporting clinical research, as I do, will agree with me. Clinicians struggling to understand statistical concepts will be satisfied.
As a biostatistician, I need textbooks and references with integrals of probability functions and mathematical proofs. I have to be able to validate the function of statistical software and to do other advanced analyses. The clinician, taking statistics as an ingredient of a well-rounded medical education, needs intuitive decision-making tools. Dr Glantz’s Primer of Biostatistics would be useful to professional statisticians in helping communicate statistical concepts to clinicians. As for clinicians at any phase of their education, from medical school through retirement, Dr Glantz’s Primer of Biostatistics has just become my number one recommendation.