The new Lehmann for the 20th century, this text presents the essential theorems of testing hypotheses while providing many original solutions for long standing problems including exact tests for two-way designs.

Unlike the full textbooks which slog through the material, Manly briskly moves along through the topics and illustrates the techniques with real numbers. References to the original papers are available as are hint for computer support -- which are sadly dated (Lotus 1-2-3?). Lightweight, easy to read on the bus, and nicely typeset.

I am an ecology grad student and I have returned to this text again and again.

At first it seemed this book was [spendy] but once I bought it I had no regrets. Mastering the material, means that one gets a glimpse at the potential exquisite simplicity of the p-adics. You also get a real sense that the proof of the greatest problem of all - the Riemann Hypothesis - is not far away if Haran's philosophy is pursued just one step further....

I absolutely appreciated this book, which, in my opinion, perfectly matches with another book by Lawrence Perko titled "Differential equations and dynamical systems". The main reason why I think this is a very good book is that it deals in a quite clear and extremely precise way with the analysis of attractors for stochastic time series, also developing non trivial concepts such as Lyapunov functions and the study of chaos. Thanks to its completeness and to a great number of examples, it can be extremely useful for both the applied mathematician who wants to improve his or her knowledge concerning the analysis and the application of both deterministic and stochastic dynamical systems and for the econometrician as well. One only fault: it should have been written later: the concept of non linear cointegration could have been dealt with, making the text even more useful for econometricians. A must read, anyway.

Analysis of repeated measurement data is commonplace in clinical trials and there is a great body of literature and books on the repeated measures analysis using linear models. The many fine texts on repeated measure linear models are often found with the term longitudinal data analysis because the repeated measurements are given over time. As Davidian and Giltinan point out htere is a need for nonlinear models in the case of pharmacokinetics trials. The theory is new and is presented here in a text for the first time. The authors provide many real applications from their experience. They present a good mix of theory and applications. Thorough references to the literature are given. Even though the development is new the authors do treat the computational aspects. Software implementation for Bayesian Hierarchical models is also mentioned. SAS macros that have been developed are also covered. The methodology is developed in the first eight chapters with applications and case studies in Chapters 9-11. Chapter 12 provides open problems to interest researchers. It contains useful information for practitioners but requires mathematical sophistication.

Albert Gifi is the pseudonym for a group of brilliant reaearchers that worked for many years on this seminal work. Chapter 1 gives an excellent overview of where ALL the roots are for current non-linear multivariate modelling. The general aim is plotting multidimensional relations in two-dimensional space, for either nominal, ordinal or continuous variables and any combination of these. Coordinates of these plots can subsequently be used as inputs for other procedures. The needed programs (provided in SPSS module caegories) are quite good and user friendly.

Along with Gallant's book on nonlinear regression this book is one of the best. It is on an intermediate level and assumes some knowledge of linear regression. Notable unique aspects of the book are (1) good use of geometry, (2) discussion of derivative free methods, (3) use of profile likelihood and profile t plots and (4) measures of degree of nonlinearity of the problem. The authors also provide many real world applications directly from their consulting experience.

These authors were among the researchers that developed these measures of nonlinearity and so that aspect of the book is original and is emphasized in this book more than in others such as Gallant's.

Gallant's book was the first clear and concise treatment of nonlinear statistical regression models. It provided general theory, examples of applications and was very well written. The book by Bates and Watts came out one year later and added measures of the degree of nonlinearity of a problem. Still the first source that I would go to when implementing a nonlinear model.

This book provides an excellent overview of chaos theory concepts applied to time series analysis. First part constitutes a good tutorial on chaos theory and its implications on time series analysis while the second part discusses in detail aspects of time-series related chaos theory concepts (with an historical perspective of the related research). Time series analysts will certainly benefit from it thanks to its balanced exposition of issues of chaos theory concepts for non-infinite data sets...

However, the only drawback is that it essentially deals with deterministic systems, not stochastic ones. But if you gathered your data on a physical system, it's OK.