Data appears to be scrambled

CPS
1

Dear IPUMS,

Thanks for supplying the world with useful data. I frequently work with the ASEC, but recently downloaded a large chunk of the monthly CPS data. I’m running into an issue where the data do not appear to be identified by YEAR, SERIAL, and PERNUM, or at the very least seem to be scrambled.

I’m attaching a picture which illustrates the issue. As you can see, individuals with the same YEAR, SERIAL, and PERNUM appear to have different ages/sex across different months. I’ve tried doing the same data import using R, and experienced the same thing. Please advise.

Best,
Zhaochen He

cps error
cps error1171×711 27.4 KB

2

Unlike for the ASEC (where month is always March), you also need to use MONTH for unique identification in the monthly CPS (as explained here).

Hope this helps!

1 Like
3

I understand, but shouldn’t data from the same SERIAL, PERNUM and YEAR be from the same person, even if they are in different months? I thought individuals were interviewed for 8 months in total. In the data, it appears that rows with the same year/serial/pernum refer to different people, not the same person interviewed at different times (see screenshot).

4

Check CPSID and CPSIDP to track households and persons across different CPS monthly samples.

Their values from your image indicate that, for example, lines 2 to 4 is the same person which is consistent with the age and gender variable.

1 Like