Thanks for your response. Am I correct that EMPSTAD is not available in IPUMS? Is there anything I can do about that?
Also, if I am able to define the sample correctly, why should I only expect results to be wthin the margin of error rather than identical to the official table?
EMPSTATD is the detailed version of EMPSTAT. If you create a data extract that includes EMPSTAT, it will automatically include the variable EMPSTATD. You can see the values of EMPSTATD by clicking “Detailed codes” on the EMPSTAT codes page.
The Public Use Microdata Sample (PUMS) are a subsample of the full (confidential) American Community Survey sample. As a result, the IPUMS data produces estimates that differ slightly (within the margin of error) from those reported in the official tables, which use the full ACS microdata.
Hope this helps.