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

Test sentivity of state marriage markets #20

Closed AaronGullickson closed 3 years ago

AaronGullickson commented 3 years ago

I need to justify sampling at the state level. A couple of options:

  1. Go whole hog and switch to a geocode that is a mix of PUMA and identifiable metro areas. This would be more work, but I am kind of interested in the option for future work anyway. I think the big problem is that I do not have these geocodes for the 1980 data because PUMAs did not exist. Instead I have "county groups" or some such.
  2. Calculate entropy measures for ethnoracial variable at national, state, and PUMA/County group level for alternate partners.
  3. Run alternate models at lower level to see if the results are consistent. What that lower level is would be tricky because limiting to metro areas may change results as well.
AaronGullickson commented 3 years ago

Ok, in general, I think the desire for comparability between the two time periods justifies the use of states, but I think I can do one sensitivity test.

  1. Create a combined geographic code which is the metro area if not missing and the state code otherwise. Then create analytical samples of 1980 and ACS data and run the basic model on each to compare parameters to the state approach.
AaronGullickson commented 3 years ago

Ok, I decided that a marriage market code that replaces state with metro area if metro area could be identified is probably the best approach in general, so I am going to just re-run everything with that coding. The server is currently chugging away at it and will hopefully be done by tomorrow.