Method to Link a National Survey to Census Tracts from Census?

[Roman Johnson]

Hi, Stata Users:

Has anyone here ever linked a national survey like the National Health and Nutrition Examination Survey or the National Health Interview Interview Survey with census tracts from the U.S. Census? I attached an article from Stimpson et al. 2007 as an example of this. Can someone provide a general method of how to do this using Stata?

Roman

[William Lisowski]

In general, you obtain your survey data with state, country, and tract codes, and then you merge it to a source of tract-level data, such as the LTDB Census Tract Data compiled at Brown University (https://s4.ad.brown.edu/projects/div...r/Bridging.htm).

The trick is getting your survey data with tract codes. Confidentiality concerns limit access to data that could possibly allow users to identify individual respondents. NHIS says the following in one of their information brochures.

To protect the privacy of survey participants, some data cannot be released on the public-use files. NHIS files containing some of these data, such as finer geographic detail, can be accessed through the NCHS Research Data Centers. Guidelines for submitting proposals to use the RDCs are available from: http://www.cdc.gov/rdc.

[Weiwen Ng]

In general, you obtain your survey data with state, country, and tract codes, and then you merge it to a source of tract-level data, such as the LTDB Census Tract Data compiled at Brown University (https://s4.ad.brown.edu/projects/div...r/Bridging.htm).

The trick is getting your survey data with tract codes. Confidentiality concerns limit access to data that could possibly allow users to identify individual respondents. NHIS says the following in one of their information brochures.

This is what I was alluding to in my earlier reply to Roman. Unfortunately, readers should note that for data collected by the US National Center for Health Statistics, there is a cost to access the restricted data (i.e. the data with geographic identifiers). In addition to that, it looks like you now have to go to one of the NCHS research data centers in person. They previously allowed remote access, and I was under the impression that this was permitted until fairly recently. The website now indicates that they don't. This seems very restrictive to me.

I don't typically do research with survey data, but if Roman is on a US campus that has a demographic research center that has access to the restricted data, there might be additional options for access. The Minnesota Population Center (I'm a PhD student at the U of M) and the Penn State Population Research Institute (applied there for PhD programs as well) are two I'm aware of, and there are others I am not aware of. If you're a student, this would probably be easier in collaboration with a professor who already has a related project going on. I vaguely recall that the MPC should have or would be able to access the geographic identifiers in the NHIS and the ACS. I don't believe we would have access to NHANES at that level (but I've never found out).