×

Regression models for categorical and limited dependent variables. (English) Zbl 0911.62055

Advanced Quantitative Techniques in the Social Sciences Series. 7. Thousand Oaks, CA: SAGE Publications. viii, 297 p. (1997).
The author says in the preface that ”This book is about regression models that are appropriate when the dependent variable is binary, ordinal, nominal, censored, truncated, or counted. I refer to these outcomes as categorical and limited dependent variables.” ”The goal of the book is to provide a unified treatment of the most useful models for categorical and limited dependent variables.” The book is based on courses on categorical data analysis taught by the author at Washington State University and at Indiana University for many years. It is assumed that the reader is familiar with basic elements of statistics that include notion, specification, estimation and interpretation of the linear regression model. In his preface the author gives a number of useful advices how to study the book.
The book contains a large variety of statistical models useful for analysing data from the area of behavioral sciences. With each model the author formulates the needed assumptions and discusses possible consequences if the assumptions are violated. The links between various models as well as problems that can arise in working with particular data sets are also discussed. Each considered model is illustrated with data from a variety of applications. Additional data sets along with sample programs are available on the author’s homepage on the World Wide Web. The reader finds also some exercises implemented in the text. The book is based on courses on categorical data analysis taught by the author at Washington State University and at Indiana University. The chapter headings are:
Introduction (1); Continuous outcomes: the linear regression model (2); Binary outcomes: the linear probability, probit, and logit models (3); Hypothesis testing and goodness of fit (4); Ordinal outcomes: ordered logit and ordered probit analysis (5); Nominal outcomes: multinomial logit and related models (6); Limited outcomes: the tobit model (7); Count outcomes: regression models for counts (8); Conclusions (9).
The book is well written. It is certainly a useful book for graduate students in social and biomedical sciences as well as for practitioners.
Reviewer: M.Huškova (Praha)

MSC:

62J05 Linear regression; mixed models
62-01 Introductory exposition (textbooks, tutorial papers, etc.) pertaining to statistics
62P25 Applications of statistics to social sciences
62J12 Generalized linear models (logistic models)
62J99 Linear inference, regression
62P15 Applications of statistics to psychology

Software:

LIMDEP; alr3
PDFBibTeX XMLCite