Hello, I’m working with one-year ACS data for New Orleans (Orleans Parish). For 2000/2005 and 2012/2013, the population totals closely match the tables by Census. But for the five years immediately following Katrina, the totals calculated from the IPUMS microdata are signficantly lower than the published populations (on the order of 25-40% smaller). National totals come out correctly. Any idea of the source of the disconnect?
I suspect this is the answer:
User Note: In the 2006-2011 ACS, persons living in Louisiana PUMAs 01801, 01802, and 01905 were all coded as living in Louisiana PUMA 77777. This is because these three PUMAs no longer had sufficient population to be included as separate entities due the effects of hurricane Katrina.
1801 and 1802 are both Orleans Parish, but 1905 is now. So I suspect the synthetic 77777 PUMA gets coded as “0” for the COUNTY variable. (Didn’t have PUMA in my extract, so now going back to confirm this is what happened.)
Assuming this is correct, is there any way to determine Orleans Parish/New Orleans residency from the IPUMS data?
Yes, the post-Katrina 77777 PUMA is coded as COUNTY=0, due to insufficient population. For the years 2006-2011, it is possible to identify respondents in Orleans Parish PUMAs 1803 and 1804, but it is unfortunately not possible to identify respondents in Orleans Parish PUMAs 1801 and 1802 using the IPUMS microdata.
Sorry that I could not deliver better news.