We are attempting to get demographic data by COUNTY and MET2013, but each of them have many missing codes.
43% missing in COUNTYand 29% missing in MET2013
There is this code book online which can map these county codes into states: https://usa.ipums.org/usa/volii/icpsrcnt.txt
However, not all codes listed. You may see there are 43% missing in this data by the table below
Here are the 4 codes don’t have states information
|
COUNTY
|
Frequency
|
|
0
|
2,300,172
|
|
8100
|
6,910
|
|
5500
|
3,077
|
|
200
|
1,789
|
Do you think you can help us with this issue?
My overall objective is to get specific demographic data that I can sort and analyze by county. Do you know of a different way to go about this?
Prompt reply is appreciated, as I am working towards a deadline early next week.
Thanks very much,
This observation is due to the fact that the lowest level geography available in public use microdata is the PUMA (public use microdata area). Although some counties are identifiable in IPUMS USA (via STATEFIP and COUNTY), counties are generally not identified in public use microdata from 1950 onward. The counties that are identifiable in IPUMS USA (from 1950 onward) are only identified because of other coterminous low-level geographic boundaries.
This is a limitation of trying to identify county-level statistics using public use microdata. I have two suggestions for a way around this limitation. First, perhaps consider performing your analysis at the PUMA level, instead of the county level. As previously mentioned, PUMAs are the lowest level geographical area identifiable in public use data. Second, if you are only interested in aggregated data, you may find the tables available via IPUMS NHGIS useful. This IPUMS project provides access to already aggregated summary tables and so the data is not subject to confidentiality restrictions that limit public use microdata.