art photo



I am trying to match data for 2006 and 2007 from the March CPS so that I can track respondents across the two years. I downloaded the NBER matching files, but they don't seem relevant to data this recent. I also read the information about matching in the CPS codebook, but it's not clear to me how to operationalize their suggestions. Would it be possible to get some assistance with this issue?


This is possible, but not all that easy to do. This is because the CPS is a survey of households so that the individuals in the household may have changed. The household is in the survey for 4 months, out for 8 months, and back in for 4 months.

The following technical documentation for the CPS describes the rotation group pattern (page 33 or so in the pdf).

The technical documentation for the March 2007 CPS describes how to match households across March 2006 and March 2007. However, once you have your matches, you need to make sure that these are the same respondents. That is a more slippery process. The description of how to match households and then the persons within the households is described on pages 14-15.

As you noted, NBER has some program code on how to do this.

If you are familiar with stata, the program examples are helpful. The readme file is also quite helpful in providing more details on which types of subject variables to use to eliminate false matches.

Let me know if there is some reason you need to use the CPS. The Survey of Program Participation (SIPP) is a CPS-like survey that is short longitudinal survey (~3 years). Thus, you might be able to look at what you want via the SIPP.

If you are specifically interested in the CPS, I can help you with this match process.

Annotated Resources:

Direct Links:

Related Question Groups: