normal (implement this, but do simulation with binomial instead. might be relevant to survey data for counties)
we're assuming that variances tau^2 are known. If they aren't, the model could be extended to incorporate this uncertainty. shouldn't change things must
true mean, sample means, sample size or variances
simulate: fixed var (n), changing sample size. and use sd units.
sigma^2/n = tau. tau is truly more of a standard error
results should be the same .
normal (implement this, but do simulation with binomial instead. might be relevant to survey data for counties)
we're assuming that variances tau^2 are known. If they aren't, the model could be extended to incorporate this uncertainty. shouldn't change things must
true mean, sample means, sample size or variances
simulate: fixed var (n), changing sample size. and use sd units.
sigma^2/n = tau. tau is truly more of a standard error