.,. the first introductory text to cover, in a single volume, models and methods for discrete dependent variables, cross-classifications, and longitudinal data. A great strength of the text is the authors' informal yet sophisticated approach, which combines the discussion of general principles with illuminating and realistic empirical examples. -Roberto Mare, University of California, Los Angeles, USA Teaching this book will be almost too easy. The prose is clear, the examples are well chosen, and the Web site provides practical details. -Michael Hout, University of California, Berkeley, USA An excellent job done by the authors. As with the first edition, Powers and Xie make the analysis of categorical data easy to understand. There are 7 chapters that are clearly written, begining with a review of simple linear regression, then going to loglinear models for contingency tables, models for ordinal and nominal dependent variables and models for event ocurrence (models for rates). The technical level is high enough to understand the theory behind the analysis and the interpretation of results. The inclusion of a new chapeter (ch.5) on multilevel models is very clearly written, and includes a short introduction to modern Bayesian modelling. Although there is not enough space for a complete introduction into this topic (which requires a high level of mathematical statistics) the authors refer to other books (like the one by Scott Lynch) for more detailed explanations (needed for a better understanding) of bayesian modelling in general. This book is a great addition to the library of students and scientists in areas like biology and sociology who want an explained compendium of (modern) techniques for analysing categorical data. Amazon review