|
SPEAKER: |
|
|
TITLE: |
"Accounting for overdispersion in the analysis of count data" |
|
DAY: |
Wednesday, February 16, 2000 |
|
TIME: |
3:30 p.m. [Coffee & cookies in BSB-202 at 3:00 p.m.] |
|
PLACE: |
BSB-108 |
The log-linear regression model is well established for relating count data to a set of explanatory variables. Frequently the residual variation from the fitted model may be larger or smaller than the expected variation under the model. After reviewing the available tests for detecting over dispersion, a generalization of these tests will be presented and their small sample properties will be discussed. The effects of overdispersion on inferences about regression parameters will also be discussed. Several environmental data sets will be used for illustration.
Abdel El-Shaarawi earned his B.S. and M.Sc. degrees from Cairo University and his Ph.D. at the University of Waterloo. He is a research scientist at the National Water Research Institute at the Canada Centre for Inland Waters in Burlington, and an Adjunct Professor in the Department of Mathematis and Statistics at McMaster. He is a founding member and President of the International Environmetrics Society and is Editor-in-Chief of its journal Environmetrics.
Dr. El-Shaarawi has served as a consultant to a large number of organizations including the National Office of Drinking Water in Morocco, WHO, IDRC (International Development Research Centre of Canada), CIDA (the Canadian International Development Agency), US-EPA, the Government of Argentina, Sydney Water Board and the Alligator River Research Institute, Australia, the Columbian Petroleum Institute, the ELF project for impact assessment of the US-Navy Communications System. Recently he served as a member of the Haystack Review Panel for evaluating NASA's current program on the assessment of risks from space debris in the Low Earth Orbit.
Dr. El-Shaarawi is an Elected Member of the International Statistical Institute, a Fellow of the American Statistical Association and the Winner of The 1996 Distinguished Medal Award of the American Statistical Association (Statistics and the Environment Section). He is also a member of the Bernoulli Society Committee on Probability and Statistics in Physical Sciences and was the Chair of its Scientific Committee for the conference held in August 1999 in Athens, Greece.
The references below have been suggested by Dr El-Shaarawi as useful background for his talk. They have been placed on reserve at Thode Library (STATS 770: Statistics Seminar).
[1] Lawless, J. F. (1987). Negative Binomial and Mixed Poisson Regression. The Canadian J. of Statistics, 15, 209-225.
[2] Prentice, R.L. (1988). Correlated Binary Regression with Covariates Specific to Each Observations. Biometrics, 44, 1033-1048.
[3] Dean, C. B. (1992). Testing for Overdispersion in Poisson and Binomial Regression Models. Journal of the American Statistical Association, 87, 451- 457.
[4] Friedl, Herwig and Stadlober, Ernst (1997). Resampling Methods in Generalized Linear Models Useful in Environmetrics. Environmetrics, 8, 441- 457.