Transportation Statistics and Microsimulation

Regression Models for Count Data    ◾    215

Example 11.2

Because overdispersion in the data of Example 11.1 seems possible, an over-

dispersed Poisson model that introduces a scale parameter to correct the

estimated covariance matrix under the Poisson model may be tried. The scale

parameter can be estimated by the square root of deviance/DF (by using SAS

PROC GENMOD with an option “scale = d”) or the square root of Pearson

chi-square/DF (by using SAS PROC GENMOD with an option “scale = p”).

The result from the overdispersed Poisson regression with the scale parameter

estimated by the square root of deviance/DF is presented in Table ...