Status:Closed    Asked:Aug 06, 2017 - 12:41 PM

Are weights necessary/"svyset" needed for regressions with NHGIS data?

I'm trying to do regression analysis between different geographic levels in the US. I've chosen NHGIS data specificlally over Ipums for this reason. Am I right in thinking that the data is pre-weighted?

If this is the case, would my command in stata need to use svyset or am I okay to treat this data as a panel?

Do you have the same question? Follow this Question

Staff Answer


Jeff Bloem


The IPUMS NHGIS project provides aggregated data in the form of tables. Thus, you are right that it does not need to be weighted like the microdata available through other IPUMS sources. Regarding treating your data as a panel: As long as your panel is of aggregated data across geographic areas over time, you can consider this data a panel. At the individual level, on the other hand, US Census data is a repeated cross-section.


Aug 07, 2017 - 11:02 AM

Report it


Why doesn't the NHGIS data include FIPS codes?
Does NHGIS have data for outside the USA?
What is the source of NHGIS data?
Using sample weights for regression analysis of Mexico's data?
Login   |   Register

Recently Active Members

View More »

Share |