Monday, November 14, 2011. 14h-16h
Paul Edwards1, Department of Biology
Regression for data that violates assumptions of simple linear regression: heteroscedastic, non-normal error structure.
The presentation focuses on count data with a single predictor variable, but time permitting we will do a logistic regression and/or a multivariate analysis.
As this is the first special topics lecture it will remain quite introductory. It assumes only a rudimentary level of comfort with R. It only requires one special package (MASS) which is installed already on all the lab computers. If you bring your own laptop we will help you install it.