NHIS Data Alert: Issues with 2015 and 2016 variables

The CDC/National Center for Health Statistics (NCHS), which produces the National Health Interview Survey (NHIS) data, has discovered several issues in the 2015 and 2016 NHIS data. We are sharing their published guidance with you and letting you know how IPUMS will approach the delivery of these data.

2016 Sample Weights

NCHS discovered a problem with how the sample was distributed between urban areas within some states. This problem has resulted in inaccurate sampling weights in the 2016 NHIS data. NCHS advises users to repeat any analyses they have conducted using the 2016 sampling weights when the revised weights are released later this year. Their investigations indicate that national estimates and estimates for key subgroups are largely unaffected when the current sampling weights are used.

IPUMS will continue to offer the original 2016 sampling weights on our website and add the revised 2016 sampling weights as soon as they are released by NCHS. Users should use these weights with caution and re-run any analyses once the new weights are available.

2015 Cancer Supplement

NCHS discovered errors in the 2015 cancer supplement data. NCHS urges users to delete the file and repeat their analyses with the revised data once they become available.

IPUMS has temporarily removed all variables that contained 2015 cancer supplement data from our website. We advise users to delete any extracts using the incorrect 2015 data. There are no problems with earlier years, but in order to immediately remove the 2015 variables we had to suppress affected variables in all years that they appear. We are working to make these variables available again soon for the period before 2015. We will add the revised 2015 cancer supplement file as soon as NCHS releases it.