A Course in Categorical Data Analysis

Preface -- Special Software -- 1 Sampling Distributions -- 1.1 Experimental design for a population proportion -- 1.2 Further properties of the binomial distribution -- 1.3 Statistical procedures for the binomial distribution -- 1.4 The Poisson distribution -- 1.5 Statistical procedures for the Poisson distribution -- 1.6 The multinomial distribution -- 1.7 Sir Ronald Fisher’s conditioning result -- 1.8 More general sampling models -- 1.9 Generalising the binomial distribution -- 1.10 The discrete exponential family of distributions -- 1.11 Generalising the multinomial distribution -- Exercises -- 2 Two-by-Two Contingency Tables -- 2.1 Conditional probability and independence -- 2.2 Independence of rows and columns -- 2.3 Investigating independence, given observational data -- 2.4 Edwards’ theorem -- 2.5 Log-contrasts and the multinomial distribution -- 2.6 The log-measure-of-association test -- 2.7 The product binomial model -- 2.8 The independent Poisson model -- 2.9 Fisher’s exact test -- 2.10 Power properties of our test procedures -- Exercises -- 3 Simpson’s Paradox and 23 Tables -- 3.1 Probability theory -- 3.2 The Cornish pixie/Irish leprechaun example -- 3.3 Interpretation of Simpson’s paradox -- 3.4 The three-directional approach -- 3.5 Measure of association analysis for 23 tables -- 3.6 Medical example -- 3.7 Testing equality for two 2 x 2 tables -- 3.8 The three-directional approach to the analysis of 23 tables (summary) -- Exercises -- 4 The Madison Drug and Alcohol Abuse Study -- 4.1 Experimental design -- 4.2 Statistical results (phase 3) of study -- 4.3 Further validation of results -- Exercises -- 5 Goodman’s Full-Rank Interaction Analysis -- 5.1 Introductory example (no totals fixed) -- 5.2 Methodological developments (no totals fixed) -- 5.3 Numerical example (a four-corners model) -- 5.4 Methodological developments (overall total fixed) -- 5.5 Business school example (overall total fixed) -- 5.6 Methodological developments (row totals fixed) -- 5.7 Advertising example (row totals fixed) -- 5.8 Testing for equality of unconditional cell probabilities -- 5.9 Analysis of Berkeley admissions data -- 5.10 Further data sets -- Exercises -- 6 Further Examples and Extensions -- 6.1 Hypertension, obesity, and alcohol consumption -- 6.2 The Bristol cervical screening data -- 6.3 The multiple sclerosis data -- 6.4 The Dundee dental health data -- Exercises -- 7 Conditional Independence Models for Two-Way Tables -- 7.1 Fixed zeroes and missing observations -- 7.2 Incomplete tables -- 7.3 Perfectly fitting further cells -- 7.4 Complete tables -- 7.5 Further data sets -- Exercises -- 8 Logistic Regression -- 8.1 Review of general methodology -- 8.2 Analysing your data using S plus -- 8.3 Analysis of the mice exposure data -- 8.4 Analysis of space shuttle failure data -- 8.5 Further data sets -- Exercises -- 9 Further Regression Models -- 9.1 Regression models for Poisson data -- 9.2 The California earthquake data -- 9.3 A generalisation of logistic regression -- 9.4 Logistic regression for matched case-control studies -- 9.5 Further data -- Exercises -- 10 Final Topics -- 10.1 Continuous random variables -- 10.2 Logistic discrimination analysis -- 10.3 Testing the slope and quadratic term -- 10.4 Extensions -- 10.5 Three-way contingency tables -- Exercises -- References -- Index.