ABSTRACT
In many areas of psychology and medical research, the investigator is interested in the number of times a fairly rare event occurs within some defined time interval. The event of interest might be the number of epileptic seizures a person has during a known period of time or the number of occurrences of a particular adverse event reported by a participant during a clinical trial. The kind of independence assumed by the Poisson distribution is extremely common in nature. The Poisson regression model uses a linear combination of factors and covariates, just like ANCOVA, but instead of using this to predict the DV as ANCOVA does, it is used to predict the log of the DV.
