You are previewing Fixed Effects Regression Methods for Longitudinal Data Using SAS.
O'Reilly logo
Fixed Effects Regression Methods for Longitudinal Data Using SAS

Book Description

Fixed Effects Regression Methods for Longitudinal Data Using SAS, written by Paul Allison, is an invaluable resource for all researchers interested in adding fixed effects regression methods to their tool kit of statistical techniques. First introduced by economists, fixed effects methods are gaining widespread use throughout the social sciences. Designed to eliminate major biases from regression models with multiple observations (usually longitudinal) for each subject (usually a person), fixed effects methods essentially offer control for all stable characteristics of the subjects, even characteristics that are difficult or impossible to measure. This straightforward and thorough text shows you how to estimate fixed effects models with several SAS procedures that are appropriate for different kinds of outcome variables. The theoretical background of each model is explained, and the models are then illustrated with detailed examples using real data. The book contains thorough discussions of the following uses of SAS procedures: PROC GLM for estimating fixed effects linear models for quantitative outcomes, PROC LOGISTIC for estimating fixed effects logistic regression models, PROC PHREG for estimating fixed effects Cox regression models for repeated event data, PROC GENMOD for estimating fixed effects Poisson regression models for count data, and PROC CALIS for estimating fixed effects structural equation models. To gain the most benefit from this book, readers should be familiar with multiple linear regression, have practical experience using multiple regression on real data, and be comfortable interpreting the output from a regression analysis. An understanding of logistic regression and Poisson regression is a plus. Some experience with SAS is helpful, but not required.

Table of Contents

  1. Praise from the Experts
  2. Copyright
  3. Acknowledgments
  4. Introduction to Fixed Effects Methods
    1. The Promise of Fixed Effects for Nonexperimental Research
    2. The Paired-Comparisons t-Test as a Fixed Effects Method
    3. Costs and Benefits of Fixed Effects Methods
    4. Why Are These Methods Called "Fixed Effects"?
    5. Fixed Effects Methods in SAS/STAT
    6. What You Need to Know
    7. Computing
  5. Fixed Effects Methods for Linear Regression
    1. Introduction
    2. Estimation with Two Observations Per Person
    3. Extending the Model
    4. Estimation with PROC GLM for More Than Two Observations Per Person
    5. Fixed Effects versus Random Effects
    6. A Hybrid Method
    7. An Example with Unbalanced Data
    8. Summary
  6. Fixed Effects Methods for Categorical Response Variables
    1. Introduction
    2. Logistic Models for Dichotomous Data with Two Observations Per Person
    3. Estimation of Logistic Models for Two or More Observations Per Person
    4. Fixed Effects versus Random Effects
    5. Subject-Specific versus Population-Averaged Coefficients
    6. A Hybrid Model
    7. Fixed Effects Methods for Multinomial Responses
    8. Summary
  7. Fixed Effects Regression Methods for Count Data
    1. Introduction
    2. Poisson Models for Count Data with Two Observations Per Individual
    3. Poisson Models for Data with More Than Two Observations Per Individual
    4. Fixed Effects Negative Binomial Models for Count Data
    5. Comparison with Random Effects Models and GEE Estimation
    6. A Hybrid Approach
    7. Summary
  8. Fixed Effects Methods for Event History Analysis
    1. Introduction
    2. Cox Regression
    3. Cox Regression with Fixed Effects
    4. Some Caveats
    5. Cox Regression Using the Hybrid Method
    6. Fixed Effects Event History Methods for Nonrepeated Events
    7. Summary
  9. Linear Fixed Effects Models with PROC CALIS
    1. Introduction
    2. Random Effects as a Latent Variable Model
    3. Fixed Effects as a Latent Variable Model
    4. A Compromise between Fixed Effects and Random Effects
    5. Reciprocal Effects with Lagged Predictors
    6. Summary and Conclusion
  10. References