JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. American Statistical Association is collaborating with JSTOR to digitize, preserve and extend access to Journal of the American Statistical Association. All use subject to JSTOR Terms and Conditions Book Reviews 911 production and research. A positive by-product of this activity is the inclusion of equivalent terms in different languages. Thus each of the 2,430 defined terms is first given in English, and then the technically equivalent term is given in six other languages (French, Spanish, Dutch, German, Italian, and Russian). Indices for the six non-English languages are given for quick reference to the English entry and the English definition.The presentation of the definitions vary. Some are as brief as one or two sentences. For example, the term "outcome" has as its definition, "A term used particularly in medical studies as a synonym of response variable. It is a character whose dependency on explanatory factors is being studied." Other take up to half of a page, such as the one for "proportional hazard." I found most definitions to be accurate, although some are too narrow and do not cover the term defined completely. For example, the definition of "lack of fit" (entry 11 13) is given only as it applies to regression analysis, with no mention of its use in the broad field of goodness-of-fit procedures. Another example is the term, "paired comparison method" (entry 1535), where the definition is given solely in terms of the two-sample t test applied to paired data from the normal distribution.Most standard terms are present. However, as expected, a book of this nature will always be incomplete. For examples, in covariance analysis (entry 86), the problem of nonparallel slopes is not mentioned and, although ascertainment (entry 113) is given a full page, ascertainment bias is not even mentioned. Also, the important topic of ROC curves is not included. Further, some of the classification rationales seem unclear. For example, Ryan's test procedure (entry 1938) is presented as a method for multiple test procedure, but the important Hochberg procedure is not.No references are given. The reader will be left to pursue topics without any literature guidance. However, -this is a dictionary and not an encyclopedia; the reader who uses it for an introduction to a topic is misusing it.The major positive aspect of Elsevier's Dictionary of Biometry is that correct definitions are given and usefully elucidated as needed in mathematical terms. The trained researcher and practitioner in biometry will find it useful as a ready reference for the definition of terms. Also, the inclusion of the defined terms in non-English languages will help promote the harmonization of terms across internatio...
