Hello, I was expecting to be able to get to the county level using the 5Yr PUMS data, but found it is still only by PUMAs. Yes, I can convert the PUMAs to counties, but some counties are too small and are grouped. The tables don't work for me as I need a different poverty level than just "Below Poverty Level"
I'm looking to get poverty levels of 165% FPL, 185%FPL, and 200% FPL by race and a few other variables
Any ideas?
Glenn is correct. NHGIS and Social Explorer redistribute the published ACS Detailed Tables (i.e., Summary File) data. I don't know of published ACS tables that cross-tabulate poverty ratio by race. In…
Jonathan Schroeder said:this approach makes the assumption that there is no variation in poverty/race correspondences among counties within a PUMA.
Right, this is a very simple method for estimation. I…
Social Explorer has those data up, as does the NHGIS. What tables do you want (by number) and what are you using to get the data. There are many cases where PUMAs span more than one county, and where counties have more than one PUMA.
PUMA definitions can also change over time, where populations cross 100,000.
Thank you, Andrew, for your response. Yes, PUMAs span, so I have to combine more than I really want to.
For the Census data, I use the PUMS data with SAS. I also use MDAT
I can look at IPUMS, but wouldn't that have the same issue? I have not looked closely at NHGIS yet
Social Explorer and NHGIS have poverty ratio by race? Good to know, since I haven't been able to find that anywhere else.
Use allocation factors for estimates. If a small county's population comprises (e.g.) 30% of the containing PUMA's population, multiply the PUMA's characteristic of interest by 0.3 to get a county estimate.
Glenn is correct. NHGIS and Social Explorer redistribute the published ACS Detailed Tables (i.e., Summary File) data. I don't know of published ACS tables that cross-tabulate poverty ratio by race. In any case, NHGIS doesn't have any. I'd guess that Social Explorer doesn't either.So you could estimate county-level characteristics from PUMAs, as Glenn indicates in his other post here: "If a small county's population comprises (e.g.) 30% of the containing PUMA's population, multiply the PUMA's characteristic of interest by 0.3 to get a county estimate." However, this approach makes the assumption that there is no variation in poverty/race correspondences among counties within a PUMA. A more sophisticated approach would take account of county-level race data and poverty data (for which the ACS does provide data, just not combined in one cross-tabulation), possibly using Iterative Proportional Fitting (IPF) to make county-based estimates in a way that's constrained by both the county-level info and PUMA-level info. I don't know offhand of a good general guide for this specific application, but I think there are multiple online resources covering IPF.
Right, this is a very simple method for estimation. I should have included that caveat.
In an ideal world, PUMAs would be defined to have broadly evenly distributed characteristics. Also, all states and counties would be rectangular, and no city would cross county lines.
Dear LornaBea,
I have a program that will do this. I worked on this problem for the Commonwealth of Virginia where they had the same issue. This basic idea is to start with PUMA level data and make an adjustment based on tract level B S D C tables. This method makes an adjustment that is more advanced than adjusting using the GeoCorr factor variables. You can cross tabulate using any PUMS variable such as Race. The Poverty ratio PUMS variable can be cut at any poverty level you like. If you work for a government entity or 501(c)(3) a consult is free.
Dave Dorer
info@dorerfoundation.org
I keep to standard ACS-5 data and I do everything by county, however I do have all counties by poverty by age for every county at ( https://public.tableau.com/app/profile/tom.lahey/viz/ACS-5/Menu ) It might help a bit. You could also pull the county by race and match. It wont tell you which race for poverty rate, but give you the county profile overall. I know B17020 has poverty by race (not sure has levels), but again its standard ACS-5. I'm not a census person, but think you can rock 95% with standard ACS-5 and get close enough to whatever your looking for in zero time. Then go deep if needed (PUMS etc.). AM also almost done with tool to better search ACS-5 data at ( https://tomlaheyh.github.io/Census-ACS-5-data/ ).