Apologies for cross-posting this. I am new to posting in this group and originally put it under ACS Basics, but I believe this is the correct area for my question
We are starting to work with the 2018 1-year PUMS data. After downloading the datasets, we always check the number of weighted observations in the Person Record and cross reference it with the Census Bureau's published Annual Estimates of the Resident Population for the United States... (http://www2.census.gov/programs-surveys/popest/tables/2010-2018/national/totals/nst-est2018-01.xlsx?#). We do this by running a frequency of the ST variable weighted by the PWGTP variable. It should and typically does match for every state. This year, however, we have a discrepancy for New Hampshire.
Our weighted count from PUMS = 1,356,410
The population estimate from Census - 1,356,458
When we compare other weighted demographic characteristics from PUMS and the online ACS table generator for NH there are discrepancies across all the age groups and for sex.
Has anyone else encountered this problem?
I should mention that we tried downloading the datasets again and had the same results in both SPSS and R.
Your 2018 PUMS estimate for the total population of New Hampshire matches the Census Bureau's estimate shown in their PUMS Estimates for User Verification.
However, the Census Bureau notes in their PUMS documentation that "Estimates generated with PUMS microdata will be slightly different from the pretabulated estimates for the same characteristics published on data.census.gov. These differences are due to the fact that the PUMS files include only about two-thirds of the cases that were used to produce estimates on data.census.gov, as well as additional PUMS edits."
I hope this helps.
In reply to Mark Mather: