applying a rural/urban classification to ACS PUMA 1% files

Hi Samara -

I would take one of two approaches depending on what you mean by the urban/rural character. The USDA classifications are based on commuting patterns are reflect the proximity to large urban areas. The Census Bureau's urban/rural area are based on the rural nature of the immediate environment. A rural area near a large city could be classified as rural using the Census measure and urban using the USDA measure.

1) Proximity to urban areas: take the USDA's rural-urban commuting area (RUCA) tract-level codes, link them with the 2010 Census tract to 2010 PUMA relationship file and a census tract population file, and calculate a population-weighed RUCA value for the PUMAs. To reduce the 33 RUCA codes to a smaller set of categories, see depts.washington.edu/.../ruca-uses.php

2) Rural nature of the immediate environment: take the Census 2010 P2 table, link it with the 2010 Census tract to 2010 PUMA relationship file, and calculate the percent of the population living in rural and urban areas in the PUMA.

The 2010 Census tract to 2010 PUMA relationship file can be found at www.census.gov/.../centract_rel.html

-- Dave Stinchcomb

Dave,
Your advice has been incredibly helpful and very appreciated. I am using the second avenue in which you suggested using census tract rural/urban data to construct a rate of rurality by PUMA. However, I am confused by the tract/PUMA relationship file. When I open the census tract/2010 PUMA crosswalk file, I see that some of the PUMAs include numerous states. I'm not sure that I can paste an image of the table into this message, but I will say that PUMA 00500 (for instance) contains tracts from states 37-41, 47 and 56. This doesn't make sense and resulted in absurd situations, such as one of the PUMAs including millions of observations. Is there a state prefix for the PUMA that I should be using? Some thing else? Thanks so much for your support.

Dave,
Your advice has been incredibly helpful and very appreciated. I am using the second avenue in which you suggested using census tract rural/urban data to construct a rate of rurality by PUMA. However, I am confused by the tract/PUMA relationship file. When I open the census tract/2010 PUMA crosswalk file, I see that some of the PUMAs include numerous states. I'm not sure that I can paste an image of the table into this message, but I will say that PUMA 00500 (for instance) contains tracts from states 37-41, 47 and 56. This doesn't make sense and resulted in absurd situations, such as one of the PUMAs including millions of observations. Is there a state prefix for the PUMA that I should be using? Some thing else? Thanks so much for your support.