sum of grouped PUMA estimates within a state not equalling state estimate

Diana Lavery over 9 years ago

Hi,

I'm working with all the 3-year ACS PUMS files for the state of Texas. I have state-level estimates, and 10 sub-state/regional estimates that are groups of PUMAs. I can't get the sum of the the 10 regional estimates to equal the state-level estimate for all 3-year ACS time periods. The numbers are relatively close: for example for '05-'07 at the state-level I get 5,564589 workers in the TX labor force who meet certain criteria, and 5,347,171 for the sum of the estimates for the10 groups of PUMAs.

Am I doing something wrong, or is there some reason as to why the sums might not add up exactly (perhaps due to the way ACS samples, uses some kind of controls, etc.)?

Has anyone else ran into this before?

Thanks,
Diana

Doug Hillmer over 9 years ago

Is the state-level estimate that you're trying to match from the published 05-07 tables? If so, those estimates cannot be matched exactly by the PUMS estimates because the PUMS data is a sub-sample from the full sample which is used to create the published tables.

Doug Hillmer
Cancel
Up 0 Down

Reply

Cancel
Tom Bell over 9 years ago

If you got the results from the state totals and the region (defined as set of PUMAs) both from the public use microdata sample then I expect you have one or two PUMAs that are not assigned to region.
Cancel
Up 0 Down

Reply

Cancel
Diana Lavery over 9 years ago

Thanks so much Doug and Tom.

-Diana
Cancel
Up 0 Down

Reply

Cancel