NHIS-IPUMS and NHIS discrepancy in mortality data


I merged NHIS mortality data from NHIS-IPUMS and NHIS website for 1986-2009. IPUMS lists 266,429 deceased and 1,343,4508 presumed alive; NHIS 265,751 deceased and 1,339,495 alive. Similarly, NHIS-IPUMS lists 101,788 ineligible, whereas NHIS lists as 101,746 ineligible. Even though the difference is small, but I would like to ask the reason for it. Thanks


The discrepancy you have noted is due to the Hispanic oversample in the 1992 data (data for all other years aligns perfectly between the two sources). It appears that the values for those in the oversample were not getting merged onto the data during reformatting. The new linked NHIS-NDI data has just become available, so our team will make sure this issue is fixed when we release the updated linked mortality data, likely in the beginning of March.

In the meantime, we like to reward users who find errors in the data with an IPUMS mug. Please email your mailing address to ipums@umn.edu and we’ll have one sent to you!