1. Two Measures of Risk: Odds Ratios and Average Rates

Odds Ratio

Properties of the Odds Ratio

Three Statistical Terms

Average Rates

Geometry of an Average Rate

Proportionate Mortality “Rates”

2. Tabular Data: The 2 × *k* Table and Summarizing
2 × 2 Tables

The 2 ×

*k* Table

Independence/Homogeneity

Independence

Homogeneity

Regression

Two-Sample: Comparison of Two Mean Values

An Example: Childhood Cancer and Prenatal X-ray Exposure

Summary of the Notation for a 2 ×

*k* Table

Summarizing 2 × 2 Tables: Application of a Weighted Average

Another Summary Measure: Difference in Proportions

Confounding

3. Two Especially Useful Estimation Tools

Maximum Likelihood Estimation

Four Properties of Maximum Likelihood Estimates

Likelihood Statistics

The Statistical Properties of a Function

Application 1: Poisson Distribution

Application 2: Variance of a Logarithm of a Variable

Application 3: Variance of the Logarithm of a Count

4. Linear Logistic Regression: Discrete Data

The Simplest Logistic Regression Model: The 2 × 2 Table

The Logistic Regression Model: The 2 × 2 × 2 Table

Additive Logistic Regression Model

A Note on the Statistical Power to Identify Interaction Effects

The Logistic Regression Model: The 2 × *k* Table

The Logistic Regression Model: Multivariable Table

Goodness-of-Fit: Multivariable Table

Logistic Regression Model: The Summary Odds Ratio

Description of the WCGS Data Set

5. Logistic Regression: Continuous Data

Four Basic Properties of Multivariable Regression Model Analysis

Additivity

Confounding Influence

The Geometry of Interaction and Confounding

The Geometry of Statistical Adjustment

Logistic Regression Analysis

6. Analysis of Count Data: Poisson Regression Model

Poisson Multivariable Regression Model: Technical Description

Illustration of the Poisson Regression Model

Poisson Regression Model: Hodgkin Disease Mortality

The Simplest Poisson Regression Model: The 2 × 2 Table

Application of the Poisson Regression Model: Categorical Data

Application of the Poisson Regression Model: Count Data

Poisson Regression Example: Adjusted Perinatal Mortality Rates

First Approach: Weight-Specific Comparisons

Second Approach: A Model-Free Summary

Third Approach: Poisson Regression Model

7. Analysis of Matched Case–Control Data

The 2 × 2 Case–Control Table

Odds Ratio for Matched Data

Confidence Interval for the Matched-Pairs Odds Ratio

Evaluating an Estimated Odds Ratio

Disregarding the Matched Design

Interaction with the Matching Variable

Matched Pairs Analysis: More than One Control

Matched Analysis: Multilevel Categorical Risk Factor

Conditional Analysis of Logistic Regression Models

Conditional Logistic Analysis: Binary Risk Factor

Multiple Controls per Case

Conditional Logistic Analysis: A Bivariate Regression Model

Conditional Logistic Analysis: Interactions with the Matching Variable

Conditional Logistic Analysis: *k*-Level Category Risk Variable

Conditional Logistic Analysis: Continuous Variables

Additive Logistic Regression Model

8. Spatial Data: Estimation and Analysis

Poisson Probability Distribution: An Introduction

Nearest-Neighbor Analysis

Comparison of Cumulative Probability Distribution Functions

Randomization Test

Bootstrap Estimation

Example: Bootstrap Estimation of a Percentage Decrease

Properties of the Odds Ratio and the Logarithm of an Odds Ratio

Estimation of ABO Allele Frequencies

An Important Property (Bootstrap versus Randomization)

A Last Example: Assessment of Spatial Data

9. Classification: Three Examples

Dendogram Classification

Principal Component Summaries

Genetic Classification

A Multivariate Picture

10. Three Smoothing Techniques

Smoothing: A Simple Approach

Kernel Density Estimation

Spline Estimated Curves

Data Analysis with Spline-Estimated Curves: An Example

11. Case Study: Description and Analysis

12. Longitudinal Data Analysis

Within and Between Variability

A Simple Example

Elementary Longitudinal Models: Polynomial Models

Elementary Longitudinal Models: Spline Models

Random Intercept Model

Random Intercept and Random Slope Regression Model

Mechanics of a Variance/Covariance Array

13. Analysis of Multivariate Tables

Analysis of Ordinal Data

Wilcoxon (Mann–Whitney) Rank Sum Test

Correlation Between Ordinal Variables

Log–Linear Models: Categorical Data Analysis

Independence in a Two-Way Table

Tables with Structural Zeros

Capture/Recapture Model

Categorical Variable Analysis from Matched Pairs Data

Quasi-Independence: Association in a

*R* ×

*C* Table

The Analysis of a Three-Way Table

Complete Independence

Joint Independence

Conditional Independence

Additive Measures of Association

14. Misclassification: A Detailed Description of a Simple Case

Example: Misclassification of the Disease Status

Example: Misclassification of the Risk Factor Status

A Few Illustrations of Misclassification

Agreement Between Two Methods of Classification: Categorical Data

Disagreement

A Measurement of Accuracy: Continuous Data

Parametric Approach

Nonparametric Approach

A Detailed Example of a Nonparametric ROC Curve

Area Under the ROC Curve

Application: ROC Analysis Applied to Carotid Artery Disease Data

Agreement Between Two Methods of Measurement: Continuous Data

A Statistical Model: “Two-Measurement” Model

An Alternative Approach: Bland–Altman Analysis

Another Application of Perpendicular Least-Squares Estimation

15. Advanced Topics

Confidence Intervals

An Example of a Bivariate Confidence Region

Confidence Band

Nonparametric Regression Methods

Bivariate Loess Estimation

Two-Dimensional Kernel Estimation

Statistical Tests and a Few of Their Properties

Power of a Specific Statistical Test: The Normal Distribution Case

Power of a Statistical Test: The Chi-Square Distribution Case

Three Applications

Multiple Statistical Tests: Accumulation of Type I Errors

References

Index