1

I am reviewing the tables at https://data.census.gov/cedsci/

But it seems like they are catering to people that want some columns or some locations/years.

But for my project I am looking to get every year available (I think ACS goes back to 1990 at PUMA level), for every PUMA, for every variable available.

This will allow me to run year-level regressions for the whole US with the high precision of geography.

1

The American Community Survey (ACS) began large-scale surveying starting in 2005, and began nationally-representative population surveying in 2007 (when it included group quarters populations). Prior to that, the ACS was inherited from the Decennial Census Long Form. Essentially, your data years would look something like this: 1990, 2000, 2007-2018. The microdata are available in the Public Use Microdata Sample (PUMS) series. These PUMS have substate geographic delineations known as PUMS Areas (PUMA's) that are divided to cover a minimum population threshold for each PUMA. As a result, PUMA boundaries change every 10 years and are often not comparable to each other between Decennial Censuses. This may greatly affect your ability to compare 2010 with 2000 or 1990 if your goal is substate analysis, for instance.

If you wish, you can find most of the non-spatial data through IPUMS. They also have the geographic boundaries through their NHGIS project. Registration is free for these platforms.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.