##
**Multilevel statistical models.
3rd ed.**
*(English)*
Zbl 1014.62126

Kendall’s Library of Statistics. 3. London: Arnold. xii, 253 p. (2003).

Many important research questions arising in the social and biolocical sciences concern hierarchical structures. Units of primary interest fall naturally (and nonrandomly) into groups that may themselves be regarded as units of analysis at the next higher level of aggregation. One example occurs in educational research, where individual students are nested in classes that are in turn grouped in schools.

The present volume aims to provide a comprehensive coverage of multilevel models. An introduction to multilevel models is given first. Then, the basic two-level models and the three-level models and more complex hierarchical structures are considered. Having introduced the basis of multilevel models, quite a lot of statistical methods are dealt with under the point of view of multilevel models: discrete response data, repeated measures data, multivariate data, factor analysis, nonlinear models, sample surveys, event history models, and cross-classified data structures. Additionally, the book has chapters on multiple membership models, measurement errors in multilevel models, and on missing data in multilevel models.

Methodological derivations are given in appendices. Examples and diagrams are used to illustrate the applications of the techniques. The author’s principal applications are from the social sciences. Readers should have some background in statistics. There are so many topics touched that they could not be dicussed in deep in a book of this size. For this reason, the book is more suitable as a starting point than as a general reference.

The present volume aims to provide a comprehensive coverage of multilevel models. An introduction to multilevel models is given first. Then, the basic two-level models and the three-level models and more complex hierarchical structures are considered. Having introduced the basis of multilevel models, quite a lot of statistical methods are dealt with under the point of view of multilevel models: discrete response data, repeated measures data, multivariate data, factor analysis, nonlinear models, sample surveys, event history models, and cross-classified data structures. Additionally, the book has chapters on multiple membership models, measurement errors in multilevel models, and on missing data in multilevel models.

Methodological derivations are given in appendices. Examples and diagrams are used to illustrate the applications of the techniques. The author’s principal applications are from the social sciences. Readers should have some background in statistics. There are so many topics touched that they could not be dicussed in deep in a book of this size. For this reason, the book is more suitable as a starting point than as a general reference.

Reviewer: R.Schlittgen (Hamburg)