Stroke mortality in PA: [Insert something about how stroke is the 4th leading cause of death and thus it's super important we study trends in stroke mortality...] Here, we want to look at county-level stroke mortality rates among those aged 65-74, 75-84, and 85+ from 2016 in the state of Pennsylvania. Specifically, our data consist of the number of deaths due to stroke, Yia, from county i and age-group a out of a population of size nia, where i = 1, . . . , Ns and a = 1, . . . , Ng (Ns = 67 and Ng = 3). 1. Assume Yia Pois (niaia) where ia Gam (Y0a, n0a). I claim that we can interpret Y0a and n0a as the prior number of deaths and prior population size for age a, respectively.

Question: Due to data confidentiality issues, all counts Yia < 10 have been suppressed in publicuse data. How can we account for this in our analysis?

Question: Write out the full hierarchical model; i.e., something like: p ([all of the unknown parameters] | [all of the known data]) [Likelihood] [All of the priors] .