CPS Voter Supplement Survey Design Question

Hi IPUMS Support and Forum,

I’m working with the voter supplement survey and had 2 questions, when you get a chance:

  1. The CPS uses a multi-staged stratified sampling design, but I was wondering if you could further clarify: is it a stratified 2-stage sampling design without replacement? And if so, what are the variables corresponding to each stage?

  2. Is VOSUPPWT the only weight we have to account for if we are not using variables outside of the voter supplement?

Thanks so much! Greatly appreciate any support you may be able to provide.

  1. You are correct that the CPS is based on a stratified two-stage sample design. The first stage includes selection of primary sampling units (PSU’s) and the second stage includes selection of sample housing units within each PSU. You can read more about this on page 5 of the Voting and Registration 2018 Technical Documentation and read more general information about CPS sample design on this site as well as Technical Paper 66. I assume you asked about variables in relation to variance estimation to ensure correct standard errors. For basic monthly data, WTFINL accounts for the sample design; for ASEC, REPWT (replicate weights) should be used which allows for standard errors; see this forum post.

  2. Yes, VOSUPPWT should be used for analyses using variables from the voter supplement. More information is available on the description tab for VOSUPPWT.