This document explores a few topics that we discussed during the previous meeting. The original document we discussed is available here and has some more of the basic data exploration. The additional topics included here are"

First, this histogram shows the overall distribution of the limited English proficiency (LEP) variable that we previously disucssed for all 3,142 counties. The Census Bureau defines a respondent as LEP if they are over the age of 5, speak a language other than English, and self-identify as speaking English less than “very well”. More than 90% of counties have a LEP rate of below 10%. Once again, the CAHs that score highly in one generally score highly in the other.

Additional Information about “Urban” CAHs

We previously focused primarily on CAHs in rural areas, but there are also quite a few CAHs in urban counties (using the definition of rural from the Office of Management and Budget). It’s important to think about how these might be incorporated into the project. The table below shows data for each CAH and can be filtered by whether you want to look at those in urban or rural counties. There are 265 CAHs in urban areas and 1,087 CAHs in rural areas.

Most of the urban CAHs with the populations that have high LEP are on the West Coast. Given that there are 265 CAHs considered “urban” it would be valuable to include them in the study, especially if they tend to have different translation practices than CAHs in rural counties. While we could decide to simply incorporate a few urban CAHs, another alternative is to explore how many of these CAHs are in urban counties but in ZIP codes that are considered rural. We can do this using rural-urban commuting areas codes from the USDA, which classify ZIP codes as rural or urban using population density, urbanization, and daily commuting. The table below shows each of the CAHs with mismatched county and zip codes (i.e. a rural ZIP in an urban county or vice-versa).

There are 169 CAHs in urban counties, but in rural ZIPs (64% of all CAHs in urban counties), and 38 CAHs in rural counties but urban ZIPs. Given that the majority of “urban” CAHs by the OMB definition fall into rural ZIP codes using the USDA codes, we could think about adding considering these CAHs as rural too, or at least sampling a few CAHs from this category. The table can be filtered by county or ZIP type to further explore the data.

Information about Other Languages Spoken

We are also planning to include at least some counties with high LEP where languages other than Spanish are the primary non-English language spoken. I broke down the ACS LEP data further to get some info on which languages are spoken in CAH counties. The ACS categories are a little limited, but besides English they include Spanish, other Indo-European languages, Asian and Pacific Island languages, and other languages. This is at least enough to give us a sense of what counties have large LEP populations that don’t speak Spanish.

To analyze the portion of non-Spanish-speaking LEP residents, I combined the total population estimates for each of the other language categories into a single composite measure. The table shows this information for all currently operating CAHs. It’s arranged according to the highest percent of the population with LEP that speak a non-Spanish language, and includes separate variables for the population percentage that speaks Spanish or speaks another language. It also includes a variable for the most commonly spoken “other” language.

There are a number of counties with a significant percentage of residents that don’t speak Spanish but are LEP. We could use the list from this table as a good starting point to identify CAHs that may have/need translation services for languages other than Spanish.

Additional Measures

We also talked about a couple other measures we could use to approximate the need for translation services besides percentage of the population with LEP. The two alternatives we discussed were households with LEP and residents that speak a language other than English at home. I still think the initial LEP variable is probably the best-aligned with our topic, but would be open to using the others. For now, I explored some basic things about their distributions and compared how they affected the list of CAHs generated. I can look into further questions if necessary.

LEP households

The table below compares the % of the population with LEP to the percent of households with LEP. Overall, a lower portion of households than residents have LEP. I think this is mainly due to how the variable is defined is at the household-level: as a household where no member 14 years or older speaks only English or speaks English “very well”. As a result, some individuals with LEP are irrelevant for this measure as long as a single household member speaks English “very well”. There could also be some differences between household size based on LEP status. Overall, I think it’s probably best to use the measure at the individual level, but for the most part similar counties rank highly in both.

Language Spoken at Home

The other measure of interest is the percent of the population that speaks a non-English language at home. This doesn’t necessarily imply they would require translation services (so it’s a bit less sound than using LEP), but could still be a fairly strong indicator for the need for translation services. This is also available at the household level, but is only included at the individual level here for simplicity. The table below shows the number of residents that speak another language at home is often quite a bit higher than the number with LEP (as would be expected).