Lectures on Categorical Data Analysis

1. Front Matter

   Pages i-xi

   PDF

2. The Role of Categorical Data Analysis

   • Tamás Rudas

   Pages 1-18

3. Sampling Distributions

   • Tamás Rudas

   Pages 19-40

4. Normal Approximations

   • Tamás Rudas

   Pages 41-56

5. Simple Estimation for Categorical Data

   • Tamás Rudas

   Pages 57-82

6. Basic Testing for Categorical Data

   • Tamás Rudas

   Pages 83-108

7. Association

   • Tamás Rudas

   Pages 109-136

8. Latent Classes and Exponential Families

   • Tamás Rudas

   Pages 137-155

9. Effects and Associations

   • Tamás Rudas

   Pages 157-182

10. Simpson’s Paradox

    • Tamás Rudas

    Pages 183-196

11. Log-Linear Models: Definition

    • Tamás Rudas

    Pages 197-223

12. Log-Linear Models: Interpretation

    • Tamás Rudas

    Pages 225-253

13. Log-Linear Models: Estimation

    • Tamás Rudas

    Pages 255-266

14. What’s Next?

    • Tamás Rudas

    Pages 267-275

15. Back Matter

    Pages 277-285

    PDF

This book offers a relatively self-contained presentation of the fundamental results in categorical data analysis, which plays a central role among the statistical techniques applied in the social, political and behavioral sciences, as well as in marketing and medical and biological research. The methods applied are mainly aimed at understanding the structure of associations among variables and the effects of other variables on these interactions.  A great advantage of studying categorical data analysis is that many concepts in statistics become transparent when discussed in a categorical data context, and, in many places, the book takes this opportunity to comment on general principles and methods in statistics, addressing not only the “how” but also the “why.”

Assuming minimal background in calculus, linear algebra, probability theory and statistics, the book is designed to be used in upper-undergraduate and graduate-level courses in the field and in more general statistical methodology courses, as well as a self-study resource for researchers and professionals. The book covers such key issues as: higher order interactions among categorical variables; the use of the delta-method to correctly determine asymptotic standard errors for complex quantities reported in surveys; the fundamentals of the main theories of causal analysis based on observational data; the usefulness of the odds ratio as a measure of association; and a detailed discussion of log-linear models, including graphical models.  The book contains over 200 problems, many of which may also be used as starting points for undergraduate research projects.  The material can be used by students toward a variety of goals, depending on the degree of theory or application desired.

• Association and Causation
• Biostatistics
• Categorical data analysis
• Causal analysis
• Data analysis
• Exponential family
• Graphical models
• Latent class analysis
• Log-linear models
• Markov properties
• Simpson's paradox
• Social Statistics
• Survey Analysis

• Center for Social Sciences, Hungarian Academy of Sciences, Budapest, Hungary

  Tamás Rudas

Tamás Rudas is Director General of the Centre for Social Sciences of the Hungarian Academy of Sciences, and Professor in the Department of Statistics of the Faculty of Social Sciences, Eötvös Loránd University, Budapest. He is also an Affiliate Professor in the Department of Statistics, University of Washington, Seattle. He is a Fellow of the European Academy of Sociology and Past President of the European Association of Methodology. He was Founding Dean of the Faculty of Social Sciences of the Eötvös Loránd University and has held visiting positions in several statistics departments in the US and Europe. His research deals with methods for the analysis of categorical data, including marginal models and other generalizations of the log-linear model, and the assessment of model fit.

Policies and ethics