search

American-Institutes-for-Research / WeMix

WeMix public repository
GNU General Public License v2.0
10 stars 2 forks source link

Using WeMix to analyze complex survey data #13

Closed awcm0n closed 10 months ago

awcm0n commented 10 months ago

I would like to use the WeMix package to fit mixed-effects models to complex survey data [specifically, data from the Medical Expenditure Panel Survey (MEPS)]. However, I'm not sure how to include complex multistage design elements—sampling strata, primary sampling units (PSU), and estimation weights, which are commonly used to implement a Taylor-series approach to estimate standard errors for weighted survey estimates—in the mix model specification. Do you have any advice for using the WeMix package with complex survey data? Thanks.

pdbailey0 commented 10 months ago

@awcm0n this is what WeMix is for, analysis of multi-stage surveys that account for sampling weights! I don't know the structure of MEPS but you can look at the vignette in WeMix for an example of a two-stage sample being analyzed with WeMix. When WeMix does this, it calculates a CR-0 covariance matrix based that accounts for arbitrary covariance between units in the groups.

If you have specific questions inside of that, feel free to ask them.

pdbailey0 commented 10 months ago

It seems there were not follow up questions, so I'm closing this.