I am working on Dade County (FL) data for 1990 5%, but I do not see data at that level for 2000 (5%) and 2010 (5%). When I ran a regression in SPSS that was filtered down to Dade level data I do not receive any data in the output. And when I check the county list I do not see Dade marked off for 2000-onward. I have spoken with a colleague who has previously used this 2000 data. Is there no data for Dade county for this data set? Or am I pulling this data incorrectly, and if so how would I correct to pull the county level data for Dade. Thank you for your help.
The reason it appears that there is no data available for Dade County, Florida after 1990 is because of the way the 2000 PUMA boundaries were drawn. In 1990 the component PUMAs for Dade County perfectly matched up with the County boundaries (IPUMS refers to this as being identifiable). In 2000 the PUMA boundaries changed and Dade County was no longer identifiable. The following are the PUMA codes that identify parts of Miami-Dade County:
PUMA, county, name, Pop. 2000, Allocation factor
0400112086Miami-Dade FL102123 1.000
0400212086Miami-Dade FL104643 1.000
0400312086Miami-Dade FL105305 1.000
0400412086Miami-Dade FL118445 1.000
0400512086Miami-Dade FL137551 1.000
0400612086Miami-Dade FL101338 1.000
0400712086Miami-Dade FL128514 1.000
0400812086Miami-Dade FL125727 1.000
0400912086Miami-Dade FL101190 1.000
0401012086Miami-Dade FL110054 1.000
0401112086Miami-Dade FL158650 1.000
0401212086Miami-Dade FL138722 1.000
0401312086Miami-Dade FL108940 1.000
0401412086Miami-Dade FL111491 1.000
0401512086Miami-Dade FL128513 1.000
0401612086Miami-Dade FL104501 1.000
0401712086Miami-Dade FL117082 1.000
0401812086Miami-Dade FL102653 1.000
0401912086Miami-Dade FL120654 1.000
0402012086Miami-Dade FL 27266 0.255
12087Monroe FL 79589 0.745
As you can see, the last component PUMA (04020) is only partially covering Miami-Dade (allocation factor of .255) but mostly covers Monroe (0.745). You could choose to only use the PUMA codes that have allocation factors of 1.0, but this would mean eliminating 27266 people in 2000. You could also choose to group Dade and Monroe County because together they are 100% identifiable. However, in 1990 Monroe is grouped with Collier County, so to identify 100% of Dade County and make your geographic representations comparable over time (from 1990 to 2010) you would have to group Dade, Monroe, and Collier County.
The variable CONSPUMA (Consistant PUMA) makes county groups identifiable from 1980 to 2011. Because CONSPUMA, stretches back to 1980 the CONSPUMA code that you would be interested in (75) groups Dade County with Monroe, Collier, Charlotte, DeSoto, Glades, Hardee, Hendry, Highlands, Indian River, Okeechobee, St. Lucie, and Martin.
I hope this helps.