Azen R., Walker C.M. Categorical Data Analysis for the Behavioral and Social Sciences

Taylor and Francis Group, 2011. — 296 p. — ISBN: 978-1-84872-836-3.Featuring a practical approach with numerous examples, this book focuses on helping the reader develop a conceptual, rather than technical, understanding of categorical methods, making it a much more accessible text than others on the market. The authors cover common categorical analyses and emphasize specific research questions that can be addressed by each analytic procedure so that readers are able to address the research questions they wish to answer. To achieve this goal, the authors:1) Review the theoretical implications and assumptions underlying each of the procedures.
2) Present each concept in general terms and illustrate each with a practical example.
3) Demonstrate the analyses using SPSS and SAS and show the interpretation of the results provided by these programs.A "Look Ahead" section at the beginning of each chapter provides an overview of the material covered so that the reader knows what to expect. This is followed by one or more research questions that can be addressed using the procedure(s) covered in the chapter. A theoretical presentation of the material is provided and illustrated using realistic examples from the behavioral and social sciences. To further enhance accessibility, the new procedures introduced in the book are explicitly related to analytic procedures covered in earlier statistics courses, such as ANOVA and linear regression. Throughout each chapter the authors use practical examples to demonstrate how to obtain and interpret statistical output in both SPSS and SAS. Their emphasis on the relationship between the initial research question, the use of the software to carry out the analysis, and the interpretation of the output as it relates to the initial research question, allows readers to easily apply the material to their own research. The data sets for executing chapter examples using SAS Version 9.1.3 and/or IBM SPSS Version 18 are available on a book specific web site. These data sets and syntax allow readers to quickly run the programs and obtain the appropriate output. The book also includes both conceptual and analytic end-of-chapter exercises to assist instructors and students in evaluating the understanding of the material covered in each chapter.
This book covers the most commonly used categorical data analysis procedures. It is written for those without an extensive mathematical background, and is ideal for graduate courses in categorical data analysis or cross-classified data analysis taught in departments of psychology, human development & family studies, sociology, education, and business. Researchers in these disciplines interested in applying these procedures to their own research will appreciate this book’s accessible approach.Introduction and overview.
What is categorical data analysis?
Scales of measurement.
A brief history of categorical methods.
Organization of this book.
Probability distributions.
Probability distributions for categorical variables.
Frequency distribution tables for discrete variables.
The hypergeometric distribution.
The Bernoulli distribution.
The binomial distribution.
The multinomial distribution.
The Poisson distribution.
Proportions, estimation, and goodness-of-fit.
Maximum likelihood estimation: a single proportion.
Hypothesis testing for a single proportion.
Confidence intervals for a single proportion.
Goodness-of-fit: comparing distributions for a single discrete variable.
Computer output: single proportion example.
Computer output: goodness-of-fit example.
Association between two categorical variables.
Contingency tables for two categorical variables.
Independence.
Odds ratio.
Testing the association between two categorical variables.
Applications: inter-rater agreement.
Computer output: test of independence.
Computer output: inter-rater agreement.
Complete example.
Association between three categorical variables.
Contingency tables for three categorical variables.
Marginal and conditional independence.
Inferential statistics for three-way tables.
Computer output: test of association.
Computer output: inferential tests for three-way tables.
Complete example.
Modeling and the generalized linear model.
Components of a GLM.
Parameter estimation.
Model fit.
Parameter inference.
Log-linear models.
Coding of variables and notation in log-linear models.
Log-linear models for two-way contingency tables.
Log-linear models for three-way contingency tables.
Inference for log-linear models: model fit.
Log-linear models for higher tables.
Computer output: fitting log-linear models.
Complete example.
Logistic regression with continuous predictors.
The logistic regression model.
Parameter interpretation.
Inference.
Model fit, residuals, and diagnostics of fit.
Complete example and computing.
Logistic regression with categorical predictors.
Coding categorical predictors.
Parameter interpretation.
Inferential procedures.
Model testing.
Interactions.
The relationship between logistic regression and log-linear models.
Computing.
Complete example of logistic regression using both categorical and continuous predictors.
Logistic regression for multicategory outcomes.
The multicategory logistic regression model.
Parameter testing and interpretation.
Multicategory logistic regression with ordinal outcome variables.
Computing and complete example using a nominal outcome variable.
Computing and complete example using an ordinal outcome variable.