By Donald Mainland. 2nd ed. 381 pp. $9. W. B. Saunders Co., W Washington Sq, Philadelphia 5, 1963
This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.
This new edition has been extensively reorganized and rewritten, but the main theme and basic feature of the book remain the same. It is not an ordinary textbook with a collection of statistical recipes. Dr. Mainland is addressing mature clinical researchers and their statistical colleagues, emphasizing good planning of an experiment rather than numerical juggling of the results. He is particularly critical of unnecessary and frequently unwarranted "significance" tests which abound in medical literature. The reader must not think that Dr. Mainland is condemning the test; he is merely against the wrong use of it.
Dr. Mainland, a strong advocate for experimentation, regards survey results as better for administrative use than for scientific investigation. He is rightly skeptical of the post hoc interpretations of the usual survey data. Hundreds of pitfalls in interpreting experimental and survey results have been illustrated by anecdotes, real and hypothetical. Many passages actually read as
Li CC. Elementary Medical Statistics.. JAMA. 1963;186(11):1027. doi:10.1001/jama.1963.03710110079028