Using Strata and Cluster variables to obtain confidence intervals vs. using replicate weights


Can you use the Strata and Cluster variables, paired with linearization, to obtain confidence intervals around estimates in the 2013 ACS?

Are there any disadvantages to doing that, as opposed to estimating confidence intervals with the replicate weights? I notice that there’s some description of this on this page (…)

which gives commands like this:

 svyset cluster [pweight=perwt], /// 

 strata(strata) svy, subpop(if age \>= 65): mean var1 


I checked out this document 


But I wasn't sure if the same findings 

(about linearization vs. replication methods) applied to the 2013 ACS.


Many thanks.


Yes, you can use the Strata and Cluster variables to obtain confidence intervals for the 2013 ACS. This paper offers some insight into different methods of calculating standard error with IPUMS data products and how the outcomes may vary. Replicate weights are said to provide more robust results, but they also require more computing time. Regardless, the confidence intervals calculated through linearization or the use of replicate weights should be comparable. In the end it is up to the researcher to decide which they feel more comfortable using.

I hope this helps.