[These are some notes intended for students in my undergraduate Introduction to Social Demography class, for use in working on their final projects, but I thought they would be of wider interest to others using IPUMS in their teaching.]
Many students are interested in doing detailed comparisons of the social and demographic characteristics of specific ethnic groups. In reviewing the project proposals, I saw that many students had used, or were planning to use, the detailed codes for the RACE variable. I would strongly encourage everyone who is interested in a specific ethnic group to assess whether some of the other available variables like the self-reported ethnicity variables (ANCESTR1 and ANCESTR2) available in 1980, 1990, and 2000 might offer more cases and better resolution.
If you want to distinguish between 1st generation and later generation, you can filter on BPL, as described below. Until 1970, a variable for father’s birthplace, FBPL, is also available, and as described below, can be used to identify the second generation.
· First-generation immigrants (Born abroad)
o Filter based on the birthplace variable, BPL. Codes for BPL identify the state or country of birth of the respondent. If you want to restrict to people born in a particular country, look up the code for that country, and use that code in the filter. For, example, people born in Sweden would be identified by BPL being equal to 405, so in the filter field on the screen for specifying your tabulation you would enter bpl(40500) along with whatever filters are relevant to your calculation.
o BPL codes: http://usa.ipums.org/usa-action/codes.do?mnemonic=BPL. Make sure to choose ‘Detailed’ rather than ‘General’ so that you see the 4 or 5 digit codes that you will use in your filters. If you use the 3 digit codes listed under the ‘General’ view, the filter will not work properly.
o With the detailed codes that you will use in your filter, there may be multiple codes corresponding to the same country, because there are different codes for regions in the same country, especially if during the nineteenth century, these regions were separate countries. For example, Canada is 15000-15083, Germany is 45300-45362.
· The second generation (born in the U.S. to a parent who was born abroad)
o You can identify people with a parent born abroad by use of the Father’s Birthplace (FBPL) or Mother’s Birthplace variable (MBPL). To make things consistent, please base your definition of the ‘second generation’ on the father’s birthplace (FBPL).
o To ensure that you are considering individuals born in the United States to father or mother who was born abroad, combine a filter based on father’s birthplace being in the country of interest (FBPL) with a filter based on own birthplace (BPL) being the United States.
o For example, to limit your tabulation to records of second-generation Swedish-Americans, that is people born here to Swedish fathers, you would include bpl(100-12092) fbpl(40500) in your filter.
o FBPL codes: http://usa.ipums.org/usa-action/codes.do?mnemonic=FBPL
o With the detailed codes that you will use in your filter, there may be multiple codes corresponding to the same country, because there are different codes for regions in the same country, especially if during the nineteenth century, these regions were separate countries. For example, Canada is 15000-15083, Germany is 45300-45362.
o In the dataset that is available for online analysis, fbpl is only available through 1970. It isn’t provided in the 1980, 1990, and 2000 data that are available online. So your tabulations involving fbpl will normally end in 1970.
· Second and later generation, 1980-2000 (born in the U.S., but claiming an ethnicity)
o In 1980, 1990, and 2000, the Census form included a question about ethnicity. The response is in the variables ANCESTR1 (for the first response) and ANCESTR2 (for the second response). The responses, as we know from the article by Hout and Goldstein on the Irish-American population, are highly subjective. Nevertheless it does allow you to get a picture of an ethnicity that includes more than just the first- or second-generation.
o Identify the second and later generations as people who specified your ethnicity of interest in ANCESTR1, but who also indicated that they were born in the U.S.
o ANCESTR1 codes (note that values differ from the BPL and FBPL codes): http://usa.ipums.org/usa-action/codes.do?mnemonic=ANCESTR1. Make sure to choose the ‘Detailed’ view rather than the ‘General’ view so that you can see the three digit codes you will need for your filter.
o For example, to limit your tabulations to records of second- and later-generation Swedish-Americans 1980-2000, you would add bpl(100-12092) ancestr1(890-900) to your filter.
o Do not include fbpl in your filter here because in the dataset that is available for online analysis, fbpl is not included in 1980, 1990, 2000. If you filter on on ancestr1 and fbpl you will end up with no cases.
o Note that because responses on the ancestry question were open-ended, and people sometimes responded with a region in a particular country rather than the country itself, to ensure you get everyone associated with a particular country, you may need to specify a range of codes. For example, Italian includes all the codes from 510 (Italian) to 730 (Venetian), so to pick up all the people who might plausibly be claimed to be Italian, you would specify ancestr1(510-730) in your filter.
· Native-born population of the U.S.
o Include bpl(100-12092) in your filter.