|
| Français |
| Home |
| NPCDS Projects |
| Workshops & Training Events |
| Contact Us |
| Partners |
MODERN APPLIED METHODS IN BIOSTATISTICS
at the University of Toronto, Medical Science Building
August 14 -17, 2006 -- 9 AM to 5 PM
Director: Paul N Corey
Co-directors: Jamie Stafford and Wendy Lou
Lecture 2: RANDOM EFFECTS MODELS
Paul Corey
Professor, Member of Institute for Medical Sciences
Adjunct Scientist - Institute for Work and Health
Cross appointments - Faculty of Nursing - Department of Physical Therapy
The classical linear model with a continuous response variable consists
of a systematic component comprised of estimable parameters to which is
added independent and identically distributed random error. Over time
this model has been extended by having :
(1) error with non constant variance.
(2) error with several components
(3) correlated error
(4) parameters in the model that are random variables
Using examples the talk explores the consequences of incorrectly ignoring
these extensions.
Paul N Corey
Paul N Corey was Director of the Graduate Program in Biostatistics for 14 years and Associate Chair of the Department of Public Health Sciences for 3 years. Paul has extensive experience in consulting, graduate teaching and supervision and has collaborated with colleagues mainly in the area of occupational and environmental health and nutrition.