Calculating count of households that meet any one of several descriptions in PUMS

Hello - I apologize if this ends up being a relatively basic question, but I am working right now to replicate the methodology used by the Universal Service Administrative Co. (USAC) to estimate the total eligible population for USAC's Lifeline program, which provides subsidies to low-income households to help them pay for phone and internet service. USAC releases information about the total eligible population in each state, but I am hoping to derive that number for each county by replicating their process. According to USAC documentation, they define eligibility as a household that reported any of the following in their response to ACS:

ACS DESCRIPTION ACS VARIABLE ACS VALUE
Medicaid, Medical Assistance, or any kind of government-assistance plan for those with low incomes or a disability HINS4 1 -- Yes
Yearly food stamp/Supplemental Nutrition Assistance Program (SNAP) recipiency FS 1 -- Yes
Public assistance income over past 12 months (any amount) PAP 1 to 30000 -- $1 to $30000 (Rounded)
Supplemental Security Income over past 12 months (any amount) SSIP 1 to 30000 -- $1 to $30000 (Rounded)
Poverty status recode indicating household income below the 135% poverty threshold POVPIP 0:135 (inclusive)

I have been trying to use PUMS to run this calculation at the county level for my area of interest (the Kansas City metro region) but am running into a problem. I understand how to use the PUMS tool to identify the number of households that meet multiple criteria at the same time, but it is not obvious to me how to set it up so that it reports the number of households that meet any one of these five without double-counting those that meet multiple criteria. I would be grateful if someone may be able to point me towards a solution or training document that may help me understand how to process this information.

  • Hi William,

    The best way to do this might be to download the data and use it in a statistical software tool, however, there are ways to set up the analysis by creating variables, using filters, and using control variables. We wrote up something recently for our blog that outlines how we did something similar for a county data request; hope this is helpful! https://www.ncdemography.org/2022/04/07/story-recipe-finding-specific-microdata-about-your-county/

    Becky

  • Thank you! This is extremely helpful and I am going through it now to learn more about what you all did. Ultimately this is all part of a larger project related to the EBB/ACP, so that makes your post doubly useful for me. I'm very grateful for you pointing this out, and also for writing all of it up in the first place so people like me can learn more. 

  • My guess is that you're going to need to use person data to get some of those variables. I'd use a statistical software program to create a dummy variable for each person that's equal to one if any of the conditions are met. Then sum the dummy variable by serialno and merge it back into the household data. If the dummy variable is greater than or equal to one, the household is eligible.  

    I personally use SAS for this, but you should be able to do it in any statistical software package.