this is another way to do Poisson regression in SPSS. The number of deaths vary between each of ten groups. Each group had different populations, so we want to explain deaths per patient-years. We think that this death rate will very by the age of the group and whether they smoke or not. The Poisson process says that the probability of a death is a constant over time, but may be higher for some groups than others. Hence we are nesting a regression inside a Poisson model. The link function is log, so the explanatory model is death rate=exp(X). See below.