12.4. Poisson Regression for Incidence Densities

Most of the time, Poisson regression is performed when you have counts plus some measure of exposure. The next example also concerns data on melanoma cases, but includes information on exposure. Thus, you are interested in fitting a model to the log rate, or incidence densities, of melanoma exposure. This involves including an offset variable in the model.

Consider Table 12.2. The counts nhi are the number of new melanoma cases reported in 1969–1971 for white males in two areas (Gail 1978 and Koch, Imrey et al. 1985). The totals Nhi are the sizes of the estimated populations at risk; they may represent counts of people or counts of exposure units. Researchers were interested in whether the rates ...

Get Categorical Data Analysis Using The SAS® System, 2nd Edition now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.