3
$\begingroup$

The American Community Survey (ACS) Public Use Microdata Sample (PUMS) comes with weights generated with the replication weight method. There are software packages such as the survey package in R that allow you to compute standard errors using these weights. But what if I want to study ACS PUMS data using Bayesian methods and compute credible intervals? How would I do that in a way that takes into account the sampling design, which is purposefully obscured in the replication weight framework (insofar as I understand it, and I admit I have not studied it closely).

$\endgroup$
0
$\begingroup$

The R-ILNA-project has a nice R package that serve this purpose. In this project a Bayesian hierarchical model is developed that sample wights also taken into account. In the proposed Bayesian spatial smoothing model, computation is carried out using the integrated nested Laplace approximation. You can download it here

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.