Hi there,

I’m using the IPUMS Online Data Analysis System to validate the customized results generated from PUMS. While doing random quality control, I found the results at county level a little confusing.

For example, for Montgomery County,MD (FIPS 24031), while using 2006-2008 3-year ACS; row as race & filter as statefip(24), county(310). SDA provided results including: white 193,636 and a total of 826,027 which are different from the results I got from PUMS: white 579,802, total 942,838.

For other counties, such as LA in CA, the numbers would match perfectly. American Factor Founder gave the total population of 942,747 and white 577,472.

I’m not sure if I missed something here, tried a few different format of the county code and didn’t solve the problem. Also SDA wouldn’t generate results for some counties such as 24033. I’m wondering if it was because of the PUMA geography or county code?

Could you help me with it?


Remember to account for the slight change IPUMS makes to the COUNTY codes for Maryland and Nevada (see User Note). For Maryland, all FIPS codes of 009 and higher (excluding Baltimore City) are shifted down by two in the IPUMS data. For example, Montgomery County is coded as 310 in the FIPS coding scheme, but as 290 in the IPUMS sample. The population totals for COUNTY=290 match the PUMS numbers you supplied above for Montgomery County.

Similarly, Prince Georges County in Maryland (county 24033) is coded as 330 in the FIPS coding scheme; however, it is coded as 310 in the IPUMS dataset. This would explain why SDA seemed to not display Prince Georges County.

Hope this helps.