I am working with a mixed effects model to see how the prevalence of a pathogen impacts population productivity for its fish host species (multiple sampling years and host populations):
stan_lmer(fish_pop_size ~ 0 + pathogen_prevalence +
(pathogen_prevalence|population) + (1|Year),
data = data)
I’m hoping to incorporate observation error into my independent variable (pathogen_prevalence), where the observation processes differ case by case (for each host fish population and year combination). In other words, I want to change the shape of the prior distribution according to the sample size for each population per year, so that out results “rely” more heavily on beta estimates from heavily sampled populations (each year) and less so on estimates derived from populations barely sampled (each year). Is this possible??
Thanks for any help!