Longitudinal Data Analysis Using SAS

A 2-Day Seminar on Regression Analysis for Panel Data
Taught by Paul D. Allison, Ph.D. 

Read 7 reviews of this seminar. 

Panel data offer major opportunities and serious pitfalls

The most common type of longitudinal data is panel data, consisting of measurements of predictor and response variables at two or more points in time for many individuals. Such data have two major attractions: the ability to control for unobservables, and the determination of causal ordering.

However, there is also a major difficulty with panel data: repeated observations are typically correlated and this invalidates the usual assumption that observations are independent. There are four widely available methods for dealing with dependence: robust standard errors, generalized estimating equations, random effects models and fixed effects models. This seminar examines each of these methods in some detail, with an eye to discerning their relative advantages and disadvantages. Different methods are considered for quantitative outcomes, categorical outcomes, and count data outcomes.

This seminar is based in part on Paul Allison’s Fixed Effects Regression Methods for Longitudinal Data Using SAS, published by the SAS Institute in 2005.

Who should attend? 

If you need to analyze longitudinal data and have a basic statistical background, this seminar is for you. You should have a good working knowledge of the principles and practice of multiple regression, as well as elementary statistical inference. It is also helpful to have some familiarity with logistic regression. But you do not need to know matrix algebra, calculus, or likelihood theory.


Participants receive a bound manual containing detailed lecture notes (with equations and graphics), examples of computer printout, and many other useful features. This book frees participants from the distracting task of note taking.

Registration and lodging

The fee of $795.00 includes all seminar materials.

Lodging Reservation Instructions

A block of guest rooms has been reserved at the Club Quarters Hotel, 1628 Chestnut Street, Philadelphia, PA at a special rate of $112 per night. This location is about a 5 minute walk to the seminar location. In order to make reservations, call 203-905-2100 during business hours and identify yourself by using group code STA625. For guaranteed rate and availability, you must reserve your room no later than May 25, 2012.


This seminar will use SAS for the many empirical examples, but lecture notes using Stata are also available to seminar participants. No computers will be provided on site and there will be no supervised exercises. However, you are welcome to bring your own laptop and perform the distributed exercises on your own time. 

Seminar outline

Opportunities and challenges of panel data.

  1. Data requirements
    1. Control for unobservables
    2. Determining causal order
    3. Problem of dependence
    4. Software considerations
    5. Linear models
  2. Robust standard errors
    1. Generalized estimating equations
    2. Random effects models
    3. Fixed effects models
    4. Hybrid models
    5. Logistic regression models
  3. Robust standard errors
    1. Generalized estimating equations
    2. Subject-specific vs. population averaged methods
    3. Random effects models
    4. Fixed effects models
    5. Hybrid models
    6. Count data models
    7. Poisson models
    8. Negative binomial models
    9. Fixed and random effects
  4. Linear structural equation models
    1. Fixed and random effects in the SEM context
    2. Models for reciprocal causation with lagged effects

Comments from last year’s participants

Of the 25 participants who took this seminar last fall, 17 rated it as “excellent”, 7 rated it as “very good”, and 1 rated it as “good”.

Here are some comments from the participants:

“This was the clearest, most organized and easiest to understand stats course that I have ever taken. I have no doubt that what I learned will be invaluable to my work in the future. I would highly recommend this course to anyone who plans to analyze longitudinal data.”
Rachel Ward, University of Pittsburgh 

“This is a great class. Even though I don’t have a formal stats background, this class enabled me to jumpstart my understanding of longitudinal data analysis. I quickly learned various models, their strengths and limitations, and most importantly leaned how to use them. In addition, I can use this material as a starting point and refer to appropriate literature to further understand this topic. Already I can use some applications for market research.”
Vijay Raghavan, Forest Labs 

“I thought I knew enough to perform longitudinal data analysis. But, I found out that there are many things that I needed to learn. Dr. Allison was especially good at explaining different methods. I am leaving here feeling confident and knowledgeable about fixed vs. random effects. The course is well suited for beginners without much experience in longitudinal data analysis, but also a very good resource for memory refreshment and update for experienced individuals.”
Beatrice Ugiliweneza, University of Louisville

“Excellent course. Highly recommended as a concise and practical refresher on LDA methods.”
Grace Mhango, Mount Sinai Medical Center

“The course strikes an excellent balance between describing statistical methods for analysis problems in general, and specifics on how to implement the methods in a specific software package.”
Cris Price, Abt Associates

“This course is extremely practical and informative for both students and professionals who need to further their analysis skill sets in a short period of time. The course increased my confidence to return to school and immediately apply the knowledge to my current projects.”
Kelly Kenzik, University of Florida 

“As a statistician working in public health, I frequently have to work with large scale longitudinal data. This course provided me with sufficient knowledge about various methods and techniques for effective data analysis. I would recommend this course to anyone working with longitudinal data.”
Tanushree Prasad, University of Pittsburgh