I would be pretty careful with Group Quarters and their characteristics with DP.  I was working on something in West Feliciana which has Angola Prison, one of the largest if not the largest  maximum security prison in the United States with over 5,000 inmates.  In 2010 the prison was  about 75% black, in 2020 it had become about 15% black.  This was laughable on its face, but the Bureau refused to correct it and refused to say how it happened. 

• The count of the number of people in each type of GQ (there are seven) was subject to noise injection.  However, the count of the number of GQ facilities by type was not, and file exists that has the number and type of GQs by block in the whole country.  However, it appears that there are some limitations of the GQ numbers in the 2020 Census.  First, each GQ was forced to have at least one occupant.  Since these are group quarters, one does not qualify as a group.  This was a rule rule of the DP algorithm.  Also there is new Post Census Group Quarter Review, which can be triggered by a request from an authorized official.  in any state, etc.  The time for submission have just ended, but to correct an error such a filing must be made, and they will not allow characteristics (e.g. race to be changed)  But if you ant GQ data, it exists all the way down to block in terms of count and number of facilities.  Unfortunately it may or may not be accurate.

I would be pretty careful with Group Quarters and their characteristics with DP.  I was working on something in West Feliciana which has Angola Prison, one of the largest if not the largest  maximum security prison in the United States with over 5,000 inmates.  In 2010 the prison was  about 75% black, in 2020 it had become about 15% black.  This was laughable on its face, but the Bureau refused to correct it and refused to say how it happened. 

• The count of the number of people in each type of GQ (there are seven) was subject to noise injection.  However, the count of the number of GQ facilities by type was not, and file exists that has the number and type of GQs by block in the whole country.  However, it appears that there are some limitations of the GQ numbers in the 2020 Census.  First, each GQ was forced to have at least one occupant.  Since these are group quarters, one does not qualify as a group.  This was a rule rule of the DP algorithm.  Also there is new Post Census Group Quarter Review, which can be triggered by a request from an authorized official.  in any state, etc.  The time for submission have just ended, but to correct an error such a filing must be made, and they will not allow characteristics (e.g. race to be changed)  But if you ant GQ data, it exists all the way down to block in terms of count and number of facilities.  Unfortunately it may or may not be accurate.

Dear Andy,

Here is some info that I found on intercensal group quarters.  The methods are part of the Population Estimates Program (PEP).  I also emailed with the people in the GQ group. You may be familiar with this info.  Here is how GQ population estimates are updated between decennial census counts (PEP)

Group Quarters
We estimate the GQ population every year by single year of age, sex, race, Hispanic origin, and facility type. The GQ method begins with the April 1, 2020 blended base GQ population. We assume that the population in GQ remains constant throughout the decade unless we receive updated data on GQ population change. Information on change to the base GQ population comes from our annual Group Quarters Report (GQR). The GQR consists of time series data from the branches of the military, the Department of Veterans Affairs, and our state partners in the Federal-State Cooperative for Population Estimates (FSCPE). Our data providers supply data at the facility level, which allows us to aggregate to all the other estimates geographies (e.g., counties and states). We use the submitted data to calculate a year-to-year change, which we then apply to the GQ population in the estimates base.
Once we have a times series of total GQ population at the facility level, we aggregate the facility-level data to the national level and apply the April 1, 2020 blended base distribution of age, sex, race, and Hispanic origin detail by major facility type to generate estimates of the GQ population by demographic characteristics. We also apply the county distribution of age, sex, race, and Hispanic origin to the county level totals. To ensure consistency, we control the county characteristics to the national characteristics and the subcounty totals to the new county totals. Finally, we aggregate the data to the necessary levels for estimates production (e.g., three age groups for county totals production and full demographic detail for state characteristics production).

www2.census.gov/.../methods-statement-v2022.pdf

It sounds like your example of the prison population should have had the population updated in a GQR.

I emailed with the people in the Census QG department and they noted, as I was aware because I am using a PUMS based model, that the household/iinstitutional/noninstitutional breakdown is in the yearly PUMS (1 and 5 year) ACS files. That is the smallest geography that they report so there are no tract level tables for GQ 3 types, just nation and state . B26101 GROUP QUARTERS TYPE (3 TYPES) BY SEX BY AGE

It looks like my Small Area Estimation (SAE) program, which uses the most recent decennial data (2020) is the best that I can do for tracts unless I can get my hands on a summary of the GQRs, which is highly unlikely.  I'm using P5 from the PL 94-171 file.  P19 (just out this May) has GQ x age x sex for 2010 and 2020. These have tract level data.

There is B26001 GROUP QUARTERS POPULATION but it only gives the total GQ population with no institutional/noninstitutional breakdown.  I don't see any reason not to have a table with institutional/noninstitutional ACS tabla at the county or tract level. It could use the  PEP update. They can drop geographies when there is a disclosure avoidance issue.  Maybe I'll make  a suggestion in the "suggestion box."

Dave