CV4 Seems to pull P5 from PL94-171 instead of DHC

NHGIS
1

I generated a time series report for CV4. Persons by Hispanic or Latino Origin [2] by Race [5*]. It looks like it pulled table P5. GROUP QUARTERS POPULATION BY MAJOR GROUP QUARTERS TYPE from the PL data rather than P5. HISPANIC OR LATINO ORIGIN BY RACE from the DHC. Anyone else run into this issue? Any solutions?

2

Thanks for reporting this issue. As it happens, another user recently reported a problem with the 2020 data in time series table AE7, Persons by Hispanic or Latino Origin [2] by Race [6*], which includes several of the same time series as CV4. I figured out just a week ago what you found to be the case: that the 2020 data comes from table P5 in the PL94-171 dataset but should come from P5 in the DHC.

We’re planning to work on a solution soon. (Our metadata that defines the tables is correct in this case, but our process that builds the data is not operating on that metadata correctly, so we’ll need to do some development to determine where the error is occurring in our code and correct it.)

Until we’ve released corrected time series, you could get the correct data from the corresponding source table (P5) from the 2020 DHC, or, if you’d be OK using 2010 census geography for your application, you could use our geographically standardized time series (which are “Standardized to 2010”). It appears that this problem affects only our nominally integrated time series and not our geographically standardized time series.

3

Thanks! FWIW, it looks like it only affects white alone, Non Hispanic/Latino and Black/AA alone, Non Hispanic/Latino. I ended up grabbing the data directly from the census API and mushed it all together. At some point, I’ll check back and rerun it without the extra side quest to mesh data sources. Cheers!

4

We have now corrected these errors. We confirmed that the errors were limited to the 2020 census counts in nominally integrated time series tables for these two time series:

  • Not Hispanic or Latino: White alone
  • Not Hispanic or Latino: Black or African American alone

Beginning with Release 9.4 of Time Series Tables, on July 17, 2025, we had defined these series to use source data from table P5 in NHGIS dataset 2020_DHCa, 2020 Census: DHC - P & H Tables [Blocks & Larger Areas]. But an error in our process caused it to derive the data from table P5 in a different NHGIS dataset, 2020_PL94171, 2020 Census: P.L. 94-171 Redistricting Data Summary File.

In effect, the 2020 values for these time series corresponded to counts of population in, respectively, “Correctional facilities for adults” and “Juvenile facilities”.

The same errors occurred at all geographic levels in each of these four time series tables:

  • CV4. Persons by Hispanic or Latino Origin [2] by Race [5*]
  • AE7. Persons by Hispanic or Latino Origin [2] by Race [6*]
  • B10. Persons by Hispanic or Latino Origin [2] by Race [7]
  • AF0. Persons Who Are Not Hispanic or Latino by Race Combination [70]

Thanks, Steven, for reporting the issue!