Monday, 6 August 2012

Quasi-Least Squares Regression

"Quasi-Least Squares Regression" extends longitudinal methods for the analysis of correlated data.

Examples of correlated data include, but are not limited to, clustered data, repeated observations, longitudinal data, multiple dependent variables, spatial data or data from population pharmacokinetic/pharmacodynamic studies.  Extra emphasis will be given to engaging in various types of modeling projects, with class and instructor discussion regarding the most appropriate ways to select an appropriate model for a given data situation, methods of constructing a model, interpreting a model, and evaluating a model for its comparative fit.  "Quasi Least Squares Regression" will be taught at The Institute for Statistics Education at  The instructors are Dr. Joseph Hilbe and Dr. Justine Shults, and the course will be based on their forthcoming book on the subject. 
For more details please visit at

Joseph Hilbe, elected ASA Fellow and the author of "Logistic Regression Models," and also "Generalized Linear Models and Extensions" and "Generalized Estimating Equations," was, until recently, the software reviews editor for "The American Statistician."  

Justine Shults, Associate Professor of Biostatistics at the University of Pennsylvania, is a co-investigator on studies in pediatrics, nutritional epidemiology, and psychiatry.

Hilbe and Shults are co-authors of the forthcoming CRC text "Quasi Least Squares Regression."

Who Should Take This Course:
Any analyst who needs to develop, interpret or assess models for complex multivariate data structures - e.g. clustered data, repeated observations, longitudinal data, multiple dependent variables, spatial data or data from population pharmacokinetic/pharmacodynamic studies.

Course Program:

Course outline: The course is structured as follows

SESSION 1: History and Theory of QLS Regression
  • Use of QLS in biomedical research and business analytics
  • How QLS compares with some popular competing approaches, including mexed-effects models

SESSION 2: Correlation Structures for Clustered and Longitudinal Data
  • Analysis of cross-sectional and clustered data (e.g. clinical data that are collected on a family who are measured at one time point; customer satisfaction scores that are measured in different departments of a store during a one-day sale)
  • Analysis of equally-spaced longitudinal data (e.g. clinical trials with measurements collected at baseline and at 6 and 12 months; monthly hospital pricing data)
  • Analysis of unequally-sapced longitudinal data (e.g. clinical trials with unequally-spaced measurements)

SESSION 3: Analysis of Data with Multiple Sources of Correlation
  • Analysis of clustered data measured over time, such as
    • Adverse events measured on different body systems of patients at several measurement occasions
    • Montly customer satisfaction scores in several departments of a store

SESSION 4: Analysis of Model Fit
  • Choice of variables to include in the regression model, e.g. age, gender,  income, severity of disease...
  • Adequacy of assumed correlation structure
  • Sensitivity of results when compared with fitting other popular approaches such as mixed models

You will be able to ask questions and exchange comments with the instructors via a private discussion board throughout the course.   The courses take place online at in a series of 4 weekly lessons and assignments, and require about 15 hours/week.  Participate at your own convenience; there are no set times when you must be online. You have the flexibility to work a bit every day, if that is your preference, or concentrate your work in just a couple of days.

For Indian participants accepts registration for its courses at special prices in Indian Rupees through its partner, the Center for eLearning and Training (C-eLT), Pune.

For India Registration and pricing, please visit us at

Call: 020 66009116


No comments: