Hello,

I am working with the 1996, 2000, 2004 and 2008 state-level waves of the Nursing Workforce Survey, a complex survey of registered nurses in the US. I am specifically looking at those RNs with APN education. This subset constitutes roughly 10 percent of the combined sample.

Each wave comes with its respective sets of sampling weights and jackknife replicate weights. Additionally, 2008 has 100 jackknife replicate weights and the others have 20.

I am wondering whether I will need to readjust these weights before I describe the survey design to Stata 13. Some concerns are:
  • Do I need to weight the sampling weights?
  • How do I reconcile the difference in the numbers of replicate weights between 2008 and the other years? How would this affect the degrees of freedom?
  • Is it correct to specify the stages of sampling units as year, state and survey ID, in success?
  • The key variable of interest is a policy one at the state level. Thus, I need to report standard errors at the same level of aggregation. How do I explicate this in the survey design?
Thank you.