Hello,

I am using the NLSY79 for research and need advice about survey weights. The NLSY79 has cross-sectional and sampling weights for each year. There are no variables that indicate the PSU or strata. How can I correct for clustering if these variables are not provided? I can use pw [weight], but without the psu and strata the standard errors may be under/overestimated. Can I use the cross-sectional or sampling weight and use robust standard errors or bootstrapping to correct for the lack of PSU/Strata? I appreciate any help you can provide.

TL