Longitudinal Data Analysis Using SAS

A 2-Day Seminar Taught by Paul D. Allison, Ph.D. 

Read reviews of this course


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 and categorical outcomes.

This is a hands-on seminar with ample opportunities to practice the various methods.


COMPUTING

This seminar will use SAS for the many empirical examples and the exercises.  However, lecture notes and exercises using Stata are also available on request. At least one hour each day will be devoted to exercises. To optimally benefit, you should bring your own laptop with a recent version of SAS (or Stata) installed. Power outlets will be provided at each seat.

There is now a free version of SAS, called the SAS University Edition, that is available to anyone. It has everything needed to run the exercises in this course, and it will run on Windows, Mac or Linux computers. However, you do need a 64-bit machine with at least 1 GB of RAM. You also have to download and install virtualization software that is available free from third-party vendors. The SAS Studio interface runs in your browser, but you do not have to be connected to the Internet. The download and installation are a bit complicated, but well worth the time and effort.  


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.


LOCATION, FORMAT, MATERIALS

The seminar meets Friday, December 2 and Saturday, December 3 at the Courtyard Embassy Row Marriott 1600 Rhode Island Ave, NW Washington, DC 20036.

The class will meet from 9 to 5 each day with a 1-hour lunch break.
 
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 $995 includes all course materials. The early registration fee of $895 is available until November 2. 

Lodging Reservation Instructions

A room block has been arranged at the Courtyard Embassy Row Marriott 1600 Rhode Island Ave, NW Washington, DC 20036. Call Marriott Reservations at 1 (888) 236 2427 or 1 (202) 448 8004 by November 2, 2016 for the special rate of $109 per night and mention that you are part of the Statistical Horizons Meeting group.


COURSE OUTLINE

  1. Opportunities and challenges of panel data.
            a. Data requirements
            b. Control for unobservables
            c. Determining causal order
            e. Problem of dependence
            d. Software considerations
  2. Linear models
            a. Robust standard errors
            b. Generalized estimating equations
            c. Random effects models
            d. Fixed effects models
            e. Hybrid models
  3. Logistic regression models
           a. Robust standard errors
           b. Generalized estimating equations
           c. Subject-specific vs. population averaged methods
           d. Random effects models
           e. Fixed effects models
           f. Hybrid models
  4. Linear structural equation models
         a. Fixed and random effects in the SEM context
         b. Models for reciprocal causation with lagged effects

Comments from recent participants

“If you have any questions about Longitudinal Data Analysis and are overwhelmed by the internet resources or worse still by all the contradictory citations out there, then sign up for this course. Paul Allison will break it all down for you and make it look easy.”
  Grace Namirembe, Tufts University

“I left the course feeling very confident in my ability to analyze longitudinal data and make informed decisions about appropriate methods. I really enjoyed this course and learned a tremendous amount.”
  Breanne Cave, Police Foundation

“This course is taught superbly in a clear and straightforward manner to provide maximum learning. Just wonderful!”
  Debbie Barrington, Georgetown University

“This course was very good for understanding varying techniques of analyzing panel data using SAS. We were able to see results and discuss reasons why the output varied from technique to technique.”
  Alex Scrimpshire, Oklahoma Stata University