How do I use WTSUPP and WTFINL to create a monthly times series?


I would like to create a time series of college enrollment rates from the IPUMS-CPS using SCHLCOLL and the monthly data from 1989 to 2010.

I understand from the documenation that “The March CPS files include two groups of people who are not included in the production of published labor force statistics: (1) members of the armed services, and (2) members of the Hispanic oversample who were interviewed in months other than March. WTFINL and EARNWT assign these groups a value of 0. Both groups are assigned non-zero values in WTSUPP.”

So if I would like to compare the enrollment rate from say, February 2000 to March 2000, how do I exclude from the March sample (1) members of the armed services, and (2) members of the Hispanic oversample so that the enrollment rate is representative of the same universe? In other words, is there a way to create a March version of WTFINL so that March and the other months have the same sampling universe?

Thank you very much for your assistance.



IPUMS-CPS does not currently offer WTFINL for the March supplement samples. However, the variable is available in the raw CPS files (variable name A-FNLWGT) which are available from NBER. Because the raw CPS files from NBER share a sort order with the IPUMS-CPS files (as long as your extract does not select cases), you can perform as sequential merge of the NBER data onto the IPUMS-CPS data and use A-FNLWGT from the NBER files with the March supplement files.

I hope this helps.



Would it be posible to elaborate more in this answer? What do you mean with sort order?

Also, doing this the time series obtained would refer to CPS Basic Monthly files exclusively?

Thank you for your help.



“Sort order” refers to the order of the records within the file. When I say that the sort order of the NBER March Supplement data file is the same as the sort order for the IPUMS-CPS March Supplement data file, I mean that the person represented in the first record (line 1) of the NBER file is the same as the person represented in the first record of the IPUMS-CPS file. This means that you can go down the whole file matching line 1 to line 1, line 2 to line 2, line 3 to line 3, etc. After merging the NBER March Supplement variable A-FNLWGT onto the IPUMS-CPS files, all persons with A-FNLWGT==0 were not in the March Basic Monthly sample. If you drop those persons from the data, the remaining persons represent persons from the March Basic Monthly sample only, making the file comparable to all the other Supplement files so that you can compare the March Supplement variable SCHLCOLL to the April Food Security Supplement variable SCHLCOLL (or the April Basic Monthly variable SCHLCOLL) as both samples only contain individuals from that month with the same universe restrictions.
The March (aka ASEC) Supplement is special in that it includes an oversample of persons from other months, meaning that those oversampled persons do not represent the March Basic population. Furthermore, since the March Supplement file contains people who did not respond to the March Basic questionnaire, many of the Basic questionnaire items are not included in the March Supplement file. All other Supplements only include persons from the month of the supplement and the supplement files include all information from the Basic questionnaire as well.
I hope this helps.