Replication data for publication


I am using data extracts I downloaded from IPUMS USA (both Census and ACS), CPS, and others for research. As part of the publication process many journals require replication files. Is it ok to share the downloaded and merged individual level data? If not, is there a way to share the data extract links so people can download the same files used in the analysis?

I think this is a common issue, but could not find a definite answer in FAQ, although some other people have asked for specific subsets of data (e.g. or Sharing Public NHIS Data). The suggestion in the last link is actually good. Making DOI links or something similar for data extracts.


The specific data use requirements regarding redistributing IPUMS data depend on the specific IPUMS Project. In IPUMS USA, you are permitted to share the sample data to meet journal data sharing agreements, but not the full count data. In IPUMS CPS you are permitted to share the data to meet journal data sharing agreements. In IPUMS International, any redistribution of the microdata is prohibited. These specific restrictions are noted and agreed upon when users create an account for each of these projects. I agree that these details could be made easier to find. I will bring it up with our staff. In the meantime, you can always doubled-check these details by clicking on the “create an account” link on each project’s website.

Thanks for the clarification Jeff. So to ensure we are on the same page, all the individual data in the ACS 2001-2017 and 2000 Census samples are ok to share. Full count I guess are only pre-1940 I guess.

The other good option, which help users and replication is to be able to request using code (API) so that one can share the exact data request and also update it when more data becomes available. It would be great to have such a tool.