CSSS/SOC/STAT 536 Applied Logistic Regression and Log-Linear Modeling for the Social Sciences

Instructor: Kevin Quinn

Overview and Class Goals

Categorical data are the most common form of data collected in the social sciences. Examples of categorical variables include: voting decisions, occupational choice, social and political attitudes measured on "7-point" scales, marital status, and a person's decision whether to enter the workforce.

This class is designed to provide students with a basic theoretical understanding and firm practical knowledge of regression models for binary and polychotomous data as well as log-linear models for 2, 3, and multiway tables. The primary goal of the class is to provide students with a set of tools that will help them make accurate inferences when confronted with categorical data.

Announcements

If you are sitting in on this course, please sign up to audit it

Course Materials

Course Syllabus (PDF postscript)

Notes on Discrete Probability Distributions (PDF postscript)

Notes on Bayesian and Likelihood Inference (PDF postscript)

Notes on Basic Math Review PDF or postscript

Notes on Matrix Algebra PDF or postscript (These were actually written by Andrew Martin for a class of his at Washington University in St. Louis.)

Notes on Motivating Binary Response Regression Models PDF or postscript

Notes on the Newton-Raphson Algorithm for Function Optimization PDF or postscript

Notes on Interpreting Logistic Regression Coefficients PDF or postscript

Online Software Resources

The R Website

Official Documentation for R ("An Introduction to R" is especially useful)

User Contributed Documentation for R

R Examples

A very simple first R session here

Example of how to plot functions in R: plotting.R

Fitting Logistic Regression Models (Example 1)

Fitting Logistic Regression Models (Example 2)

Hypothesis Testing for Logistic Regression Models (Example 3)

Fitting and Interpreting an Ordered Logit Model (Example 4)

Fitting and Interpreting a Multinomial Logit Model (Example 5)

Fitting and Interpreting a Poisson Regression Model and a Negative Binomial Regression Model (Coming Soon) (Example 6)

Fitting and Interpreting Loglinear Models (Example 7)

Substantive Applications

Applications of Binary Logit/Probit

Applications of Ordinal Logit/Probit

Applications of Multinomial Logit/Probit

Applications of Poisson Regression

Applications of Log-Linear Models

Homework Assignments

Assignment 1 (Due before class, Oct. 18) PDF or postscript

Homework Assignment 2 (Due before class, Oct. 25)

Homework Assignment 3 (Due before class, Nov. 1)

Homework Assignment 4 (Due before class, Nov. 8)

Homework Assignment 5 (Due before class, Nov. 15)

Homework Assignment 6 (Due before class, Nov. 22)

Back to Kevin Quinn's home page