ACS Data Users Group

I just downloaded the current census response rates file for Illinois. My focus in Chicago. When I matched the response rate file to the latest ACS I did not get a match.

For example, ACS 2018 for Chicago has a tract 3201.00. The incoming response rate file has tracts 3201.01 and 3201.02.

The incoming response rate file has tracts 3301.01, 3301.02 and 3301.03; the 2018 ACS file includes 3301.00.

There are 8-9 similar mismatches, always with the response rate file having more new variants of the 2018 ACS tract structure.

Ouch! Did they do retracting of fairly settled cities for 2020? Looking around, I can find no info.

Huge implications.

Unfortunately, the Bureau released the 2020 response rates in 2020 Census tracts, not 2010 Census tracts. (Really frustrating).

A similar question was posted on the Census API developer's forum at the beginning of the response perdio. I shared some code with Stuart Lynn, out of NYC, that I use to proportionally reallocate the 2020 tracts and their response back into 2010 census tracts. He has been making those 2020 responses in 2010 geographies publicly available in his Git repository. You can access that information here: https://github.com/stuartlynn/census_2020_response_rates/tree/master/data/counts_adjusted_for_2010

How can they draw new tracts before the census count is complete? Aren't tracts supposed to be based on population?

Playing with TIGER web https://tigerweb.geo.census.gov/tigerweb/ does show differences in the Census tracts and block groups.

Nice! Thanks for sharing that Rebecca! remember in 2000 getting the new tracts at some point before PL94 data so we could do what you say, reallocate blocks back to tracts from the earlier decade for comparison, but I don't know the timetable -- I don't see 2020 yet in Tiger, do you?

For those interested -- the relationship files are here: https://www2.census.gov/geo/maps/DC2020/SR20/

FWIW, they obliquely mention these in the self-response rate technical documentation, but didn't post until March 19/20.

You can find the 2020 tract boundary files here:

www2.census.gov/.../

Yes, the Census Bureau “Response Rates Map” is tabulated for the draft 2020 Tracts layer – the draft before Tracts were finalized thru the PSAP validation process.

Find that draft layer on TIGERweb: https://tigerweb.geo.census.gov/tigerweb/ It should be the version labeled "Current vintage"… which is confusing. They used that label to avoid calling it 2020 vintage, because it was draft 2020 and not final 2020.)

IF you want to compare response rates in 2020 vs ten years ago -- then read this from Steve Romalewski at CUNY: www.gc.cuny.edu/Page-Elements/Academics-Research-Centers-Initiatives/Centers-and-Institutes/Center-for-Urban-Research/CUR-research-initiatives/Census-2020-Self-Response-Data-Questions-and-Answers

because that’s a whole complicated puzzle.

--Todd Graham

Met Council Research

Interesting. I didn't think the 2020 Census Tract boundaries were released yet. The boundaries are done through the PSAP process, which indicates that the 2020 statistical area boundaries won't be released until Feb/March 2021.

www.census.gov/.../psap.html

To go back to the original question and the follow up from Bernie, new tracts were developed under the PSAP program earlier this year using housing unit counts derived from 2010 figures plus new construction since that time. Tracts are not based on population per se.