search

AaronGullickson / panethnicity_intermar

Data for "Patterns of Panethnic Intermarriage in the United States, 1980-2018" forthcoming in Demography
MIT License
0 stars 0 forks source link

Convert marriage markets to MSA+PUMA #1

Closed AaronGullickson closed 3 years ago

AaronGullickson commented 3 years ago

Currently, I am using state as the marriage market. My tests in the marriage model suggests minimal bias here, but nonetheless it seems crude. Instead I think I should create a composite marriage market id based on MSA or PUMA. In cases where the MSA is identifiable, I would use the MSA. In cases where the MSA is not identifiable, I would use the PUMA.

I will need to check on cases that live in MSAs that are not identified by IPUMS as those cases may create problems.

AaronGullickson commented 3 years ago

Pan-Ethnicity and Intermarriage

AaronGullickson commented 3 years ago

Apparently, MeCherri and I did something like this for the African intermarriage project, so I should look at that code to get an idea of how I might do this.

AaronGullickson commented 3 years ago

I am remembering now why this is so tricky. For comparative purposes between 1980 and 2014-18, it is almost impossible. The 1980 Census did not use PUMAs but rather "county groups." It is not clear how comparable these would be over time. In addition the metarea and met2013 codes have comparability issues. I think for this reason, it makes sense to stick with states as the marriage market here. This will address most of the spatial dissimilarity issues and allow me to use the whole sample without concern about lack of comparability over time.

I am still interested in creating a synthetic marriage market for a single time period, but I think I would also have to think carefully about this because some (usually smaller) metro areas get left out of met2013 because the overlap between the metro area and the PUMAs is not tight enough. For my purposes, I may want to put some of these cases into metro areas rather than offset as a separate PUMA. The crosswalk file between met2013 and PUMA2010 codes should help.