Hi, I've been analyzing incidence rates for a time and learn to use techniques that assume poisson distribution. Nevertheless, I've seen a colleague to divide number of cases by the incidence rate value and obtain the number of persons (if for example, the rate is expressed as persons-years) and use it as a proportion. I think that this is not quite a licit procedure but want to ask what do you think.

More Francisco Fernandez's questions See All
Similar questions and discussions