Matching household income from the March CPS to people in Voting and Volunteering Supplements

Hi there,

I’m trying to tabulate voting and volunteering behaviors by income decile. To do that, I’m trying to take the household income information (hhincome) from the March survey, and spread it to the households in the September and November supplements. Right now, I’m just using the “serial” identifier to do that. Here’s my stata command:

egen hhincome2 = max(hhincome), by(serial)

I’m hoping that the “hhincome2” variable spreads the March household income information to the same households in the September and November supplements. I should mention that I’m doing this separately for the September and November supplements.

Is that all kosher? If not, how can I attach income data to the people in the September and November supplements?

If it is kosher, how should I weight my final tabulations. Any help (and patience with this newbie) would be much appreciated. Let me know if I can clarify my question, too.

Unfortunately, SERIAL cannot be used to identify households across samples. Additionally, since CPS respondents are in the sample for 4 months, leave for 8 months, and then re-enter for a final 4 months, none of the respondents can be present in both March and September/November. Thus, your current STATA command is assigning the income of one household to that of another household.

In order to attach income data to September and November households, you will need to merge the September and November IPUMS supplement files with the September and November NBER basic monthly files, respectively. Keep in mind that the basic survey includes only a categorical family income variable (hufaminc), as opposed to the continuous income variable (HHINCOME) available in the March supplement.

In order to merge your IPUMS supplement sample to its respective NBER basic sample, you must include the IPUMS variables HRHHID and HRHHID2. Then, you can merge to the respective basic monthly file using HRHHID and HRHHID2 as your unique identifiers. You can find the basic monthly datasets via NBER here and the associated command files here. You can also link households from the September supplement to the November supplement using HRHHID and HRHHID2 (see here for more information on linking CPS samples).

Hope this helps.