ST 732: Applied Longitudinal Data Analysis [SPRING 2014]

MW 1:30PM-2:45PM, 1108 SAS Hall       

Instructor: Dr. Ana-Maria Staicu [last name pronounced as "styku"]

Contact Information
Office: 5242 SAS Hall / Phone: 515-0644
Email: ana-maria_staicu [at]  ncsu [dot] edu 
Office Hours
Tuesday 3:00AM - 3:50PM
Friday 4-5PM

Teaching Assistant: Xiaoshan Li  
Office Hours: Monday 11-12PM, Thursday 10-11AM. Room SAS Hall 1101
Email: xli12 [at] ncsu [dot] edu

Syllabus [including grading policy]

Statistical software:
    SAS [on-line documentation for version 9.1.3 ].
    R will also be used; R is freely available at http://www.r-project.org/.
            Intro to R. More about R: R Lab [Courtesy of Charlie Smith] 

Announcements

Midterm Soln  ;  SAS code
Final Project Details
(NEW, March 17)

[M, Jan-6-2013] First Day of Class
[M, Jan-20-2013] Martin Luther King [NO CLASS]
[M, W March 10,14-2013] Spring Break [NO CLASS]
[M, March17] Class cancelled 
[W, April-23-2013] Last Day of Class

Lecture Notes (tentative) 

Chapter 1: Introduction, class organization, grading, course overview  

Chapter 2: Matrix review (individual study). 
Chapter 3: Review: Multivariate normal. Conditional expectation.
                  Review: Linear regression: OLS and WLS estimation

Introduction to modeling longitudinal data
Chapter 4.: Balanced design arising from single population.
                  Mean response. Variation about the mean. Sources of correlation. Common correlation models.
                  Graphical summaries
Chapter      Balanced design arising from two or more populations
                  SAS code: Exploratory tools for mean and covariance (Proc CORR, DISCRIM)
                  R code: Basic exploratory tools for mean and covariance (scatterplot matrix)

Univariate repeated measures ANOVA [classical methods]
Chapter 5.1: Introduction. Statistical model (Split Plot) for balanced, regular design.
Chapter 5.2: Questions of interest and Statistical Hypotheses
Chapter 5.3: ANalysis Of VAriance
                 R code: Split -Plot designs
                 SAS code: Proc GLM  with random/repeated statement
                 Additional reference:  Proc GLM
Chapter 5.4: Violation of covariance matrix assumption 
Chapter 5.5: Specialized within-unit hypotheses and tests

Multivariate ANOVA (Chapter 6, individual study)
Limitations of the classical methods ( Chapter 7)

General linear models for longitudinal data
Chapter 8.1: Introduction
Chapter 8.2: General models for longitudinal data
Chapter 8.3: Modeling the covariance
Chapter 8.4: Mean regression parameters estimation: Maximum Likelihood and Restricted Maximum Likelihood
                SAS code: Proc REG (OLS estimation)
Chapter 8.5: Mean regression parameters inference. Model selection approaches (LRT, AIC, BIC)
Chapter 8.6: Final Remarks: main features and limitations
                SAS code: Proc MIXED/ repeated statement
                SAS resource: Proc MIXED
                R resource: library NLME - GLS()
  
Random Coefficient Model  (Growth Curve Model). Linear Mixed Model
Chapter 9.1: Introduction
Chapter 9.2: Random coefficient model
Chapter 9.3: Inference on mean regression parameters and covariance parameters
                SAS code: Proc MIXED repeated/random statement
                R resource: library NLME - LME()
Chapter 10.1: General Linear mixed effects model
Chapter 10.2:  Inference on the regression parameters and covariance parameters
                SAS code: Proc MIXED repeated/random statement
                R code: library NLME - LME() 
Chapter 10.3: Best Linear Unbiased Prediction (BLUP) for subjects effects and individual trajectories
                SAS code: Proc MIXED prediction
Chapter 10.4: Comparing nested models for the covariance: testing whether an effect is random
                SAS code: Proc MIXED testing  [Weigth lifting study]  PressData
Chapter 10.4: Accounting for covariate information


Generalized Linear Models. Scalare responses / Longitudinal data
Chapter 11.1: Introduction
Chapter 11.2: Three-part specification of GLM
Chapter 11.3: Estimation and inference for regression parameter
                     Iterative re-weighted least squares (IRWLS)
Chapter 11.4: Illustrative examples: Logistic Regression; Log llinear regression.
                     SAS code: Proc GENMOD [Myocard Infarction]
                     R code: GLM 
                     SAS code: Proc GENMOD [Horsekicks]   
                     SAS code: Proc GENMOD [Clotting data]                SAS code: Proc MIXED testing [Weigth lifting study]
Chapter 12.1: Introduction
Chapter 12.2: Specification of marginal models
Chapter 12.3: Estimation and inference for marginal models
                  Generalized Estimating Equations
                   SAS code: Proc GENMOD [Epileptic seizures]
                   SAS code: Proc GENMOD [Respiratory illness]
Chapter 12.4: Generalized linear mixed models.
                   R code:  LME4 package
                   SAS code: Proc NLMIXED [Respiratory illness]
Chapter 12.5: Population averaged vs subject specific approaches
Chapter 12.6  Illustrations.
                   SAS code: Proc NLMIXED [Epileptic seizures]



Homeworks (tentative deadlines)

Homework 1 (Due January 22, 2014)   Solution. Soln file
Homework 2 (Due February 10, 2014) Solution. Soln file
Homework 3 (Due February 26, 2014) Solution. Soln file
Homework 4 (Due March 26 2014)     Solution. Soln file
Homework 5 (Due April 23 2014)     Solution. Soln file
Homework 5 (Due April 11, 2013)       Solution: HW5 soln. Additional files

Data sets



Courses Objectives

To introduce students to statistical models and methods for the analysis of longitudinal data, i.e. data collected
repeatedly on individuals (humans, animals, plants, samples, etc) over time (or other conditions).

Prerequisite
ST 512, Experimental Statistics for Biological Sciences II, or equivalent. Thus, students should be familiar with basic notions of probability, random variables, and statistical inference, analysis of variance, and (multiple) linear regression. Familiarity with matrix algebra is also useful. We will review matrix algebra at the beginning of the course and make considerable use of matrix notation and operations throughout. ST 512 involves the use of the SAS (Statistical Analysis System) software package; thus, students are expected to have had some exposure to the use of SAS. The course is meant to be accessible both to non-majors and majors. The underlying mathematical theory will not be stressed, and the main focus will be on concepts and applications. Please see the instructor if you have questions about the suitability of your background.

Required Text
Lecture notes prepared by Marie Davidian will be used. These may be purchased at the Sir Speedy across the street from Patterson on Hillsborough. You should obtain a copy.

Useful Links: