Modelling Forest Systems

Biometrics 61, 1129–1140 December 2005 BOOK REVIEWS EDITOR: I. PIGEOT Modelling Forest Systems (A. Amaro, D. Reed, and P. Soares, eds) Jerry K. Vanclay Ranked Set Sampling. Theory and Applications (Z. Chen, Z. Bai, and B. K. Sinha) Mohammad Fraiwan Al-Saleh Multiple Analyses in Clinical Trials. Fundamentals for Investigators (L. A. Moye) Joachim Roehmel Association Schemes—Designed Experiments, Algebra and Combinatorics (R. A. Bailey) Aloke Dey Nonparametric Statistical Methods for Complete and Censored Data (M. M. Desu and D. Raghavarao) Robert F. Woolson The Statistical Evaluation of Medical Tests for Classification and Prediction (M. S. Pepe) James A. Hanley Linear Models with R (J. J. Faraway) Ronja Foraita Statistical Analysis and Data Display. An Intermediate Course with Examples in S-PLUS, R, and SAS (R. M. Heiberger and B. Holland) John S. J. Hsu Brief Reports by the Editor Testing Statistical Hypotheses, 3rd edition (E. L. Lehmann and J. P. Romano) Recent Advances in Quantitative Methods in Cancer and Human Health Risk Assessment (L. Edler and C. P. Kitsos, eds) Step-by-Step Basic Statistics Using SAS (Student Guide and Exercises) (L. Hatcher) Ronald Cody Handbook of Statistics: Data Mining and Data Visualization (C. R. Rao, E. J. Wegman, and J. L. Solka, eds) Applied Spatial Statistics for Public Health Data (L. A. Waller and C. A. Gotway) Olaf Berke Selection Bias and Covariate Imbalances in Randomized Clinical Trials (V. W. Berger) Constrained Statistical Inference. Inequality, Order and Shape Restrictions (M. J. Silvapulle and P. K. Sen) Anthony Hayter American Journal of Mathematical and Management Sciences, Vol. 23 (3&4) (E. J. Dudewicz, B. L. Golden, and Z. Govindrajulu, eds) Jan Beirlant Numerical Methods for Functions (C. G. Small and J. Wang) Nonlinear Estimating Richard Morton AMARO, A., REED, D., and SOARES, P. (eds) Modelling Forest Systems. CABI Publishing, Wallingford, U.K., 2003. 432 pp. US$140.00/£75.00 (hardcover), ISBN 0-85199-693-0. This book contains 34 papers from a workshop held in Sesimbra, Portugal in 2002. The papers are grouped into five sections: modeling strategies; mathematical approaches; estimation processes; validation; and metadata. Each section con- Analysis of Clinical Trials Using SAS: A Practical Guide (A. Dmitrienko, G. Molenberghs, C. Chuang-Stein, and W. Offen, eds) Genetic Analysis of Complex Traits Using SAS (A. M. Saxton, ed) A Handbook of Statistical Analyses Using Stata, 3rd edition (S. Rabe-Hasketh and B. Everitt) tains several papers that illustrate the diversity of approaches entertained in forest modeling. The book is not a recipe book showing “how to build a model,” but rather offers a comprehensive overview of the state of play in forest modeling, covering a broad range of issues from broad scale mapping site index to the challenge of archiving models and their metadata for future reference. This diversity makes the book a good resource to provoke discussion among graduate students. For 1129 1130 Biometrics, December 2005 instance, the first chapter by Harold Burkhart dwells somewhat on the importance of parsimony, but several subsequent chapters present models with not just tens, but scores of estimated parameters. Several papers are reflective, commenting on the development and evolution of methods and models (e.g., chapter 8 by Ralph Amateis), while others are more forward looking, outlining plans and recommendations for future work (e.g., chapter 7 by Heyns Kotze). Some papers document traditional growth and yield models (e.g., chapter 9 by Paula Soares and Margarida Tome, relating to pulpwood production) while others focus on a broader range of goods and services (e.g., chapter 26 by Paul van Gardingen). Many of the papers deal with single-species plantations, but some address natural forests with many species (e.g., chapter 21 by Nicolas Picard, Sylvie Gourlet-Fleury, and Plinio Sist relating to tropical rainforest) and another deals with landscape-scale visualization (chapter 30 by Falk-Juri Knauft). The book concludes with a series of papers examining the quality of data and models (e.g., chapter 31 by David Reed and Elizabeth Jones) and examining our efforts to conserve and document models so that they remain available for future researchers to examine and learn from (chapter 32 by Keith Rennolls). The book will be a useful resource for institutions offering PhD programs or graduate-level courses in environmental modeling. Jerry K. Vanclay School of Environmental Science and Management Southern Cross University Lismore, Australia CHEN, Z., BAI, Z., and SINHA, B. K. Ranked Set Sampling. Theory and Applications. Springer, New York, 2004. xii + 224 pp. US$59.95, ISBN 0-387-40263-2. Ranked set sampling (RSS) was introduced by McIntyre in 1952. In its simplest form, the RSS procedure consists of drawing k simple random samples (SRS) of size k each from the population and ranking the elements within each sample, by judgment, with respect to the characteristic of interest. Then the ith smallest observation from the ith sample is chosen for quantification. This procedure increases the chance of obtaining a more representative sample. This book, written by three experts in RSS, is the first that addresses most developments of RSS. A considerable part of it is based on the authors’ own research. A description and a motivation of RSS are given in chapter 1. Chapters 2–4 deal with inference; parametric as well as nonparametric inference is considered with balanced and unbalanced RSS. Results of papers of some other authors are briefly outlined, but the authors do not give enough attention to one important main result of Takahasi and Wakimoto (1968). Distributionfree tests are dealt with in chapter 5. Power comparisons are presented for the two cases of perfect and imperfect ranking. RSS based on concomitant variables is discussed in chapter 6. The multi-layer RSS, a generalization of the bivariate RSS developed earlier, as well as adaptive RSS, are outlined. In chapter 7, which I consider the most interesting, the authors explore possibilities of using the so-called repeated RSS procedure, another name for multi-stage RSS, for data reduction especially in the case of huge data sets. Six case studies are the content of the last chapter. Analysis of these data shows the advantage of the RSS procedure over SRS. The aim of the book is to give a systematic account of the theory and application of RSS; the intention was to cover each development of RSS since its birth. The object is somehow (but not completely) met; there are some developments in RSS that are either not mentioned or only cited with not enough details: e.g., the work on Bayesian RSS, and RSS and Monte Carlo integration; some variations and other applications of RSS are missing. However, this is not a serious drawback, given the huge number of papers on the topic. The book is well laid out with concepts well explained. Each chapter (except 6 and 7) ends with a historical note or a bibliography, in which the authors cite some related references. Very few typos or minor inaccuracies have been found by readers and fed back to the authors. The style is readable, with some examples throughout the book, but no exercises; this does not make it an ideal textbook. The lack of author index makes it hard to locate specific references in the text. Besides for applied and theoretical statisticians, the book can be useful for users of statistics in various fields such as agricultural, environmental, and medical sciences, etc. Those who have recently become interested in the topic will find it an excellent start. In conclusion, I recommend this book as a reference book for researchers, for practitioners, and as a textbook for graduate/special topics course. Reference Takahasi, K. and Wakimoto, K. (1968). On unbiased estimates of the population mean based on the sample stratified by means of ordering. Annals of the Institute of Statistical Mathematics 20, 1–30. Mohammad Fraiwan Al-Saleh Department of Statistics Yarmouk University Irbid, Jordan MOYE, L. A. Multiple Analyses in Clinical Trials. Fundamentals for Investigators. Springer, New York, 2003. xxiii + 436 pp. US$79.95, ISBN 0-387-00727. In his introduction the author describes the target group of readers as clinical investigators at all levels, research groups within the pharmaceutical industry, medical students, public health students, health care researchers, physician-scientists, and regulators at the local, state, and federal level. This should not preclude biostatisticians at all levels from reading this book, because there is a lot to learn for professional biostatisticians, e.g., from the excellent examples, from the analytical skills to make severe problems in some of the studies quite obvious, from the problems sections that come with each of the 13 chapters, and from the sometimes surprisingly fresh ideas and suggestions on how to cope with multiplicity. Indeed, the book tries successfully to avoid complex mathematical delineations, but this is by no means a drawback. On the contrary, it allows readers to get through the pages fluently without the need to stop reading frequently to make