Status:Closed    Asked:Sep 26, 2015 - 07:07 PM

Matching monthly CPS surveys to create a panel dataset


I would like to create a panel dataset using the monthly CPS surveys in order to track changes in labor force status. I am unsure how to do this. I consulted the paper by Drew, Flood and Warren and I think that I should be able to just cpsidp to track individuals across surveys, right? Then, is it just a matter of appending the monthly CPS surveys and sorting by cpsidp?

Any help is appreciated!

Do you have the same question? Follow this Question

Staff Answer




Technically, CPSIDP links the respondent appearing on a designated household roster line number (LINENO) across samples by assigning a unique CPSIDP value to this line number. Since the original person may leave the household, the respondent appearing on a given line number can be replaced; however, the CPSIDP value will remain constant for a maximum of 8 months (i.e. the length of time a household can remain in the CPS survey). As a result, it is important to verify your CPSIDP linkages with the AGE, SEX, and RACE variables, allowing for AGE to increase. Due to the complexity of linking persons and households across CPS samples (e.g. migration, mortality, non-response, and recording errors), erroneous links can still sometimes occur among those that match based on CPSIDP and demographic characteristics. Since the IPUMS-CPS linking variables are built using the original CPS variables, any errors in the source data will be reflected in CPSIDP. Ultimately, it is up to the individual researcher to determine the acceptability of the linkages made using CPSIDP.

Hope this helps.


Sep 28, 2015 - 11:58 AM

Report it


Pool or merge monthly CPS surveys to find yearly rates?
Login   |   Register

Recently Active Members

View More »

Share |