I am trying to merge CEPR CPS ORG data with IPUMS CPS for 2007 using the variables “statecensus year month hrhhid2 hrhhid lineno sex age” to get few additional employment variables for individuals from IPUMS CPS merged to the CEPR ORG data.

Unfortunately, my merge using the above variables show that there are no matched records between the two files. Is this possible? One would expect the respondents in the CPS ORG to have at least one record in the IPUMS CPS.

Could you please help me understand how I can merge these two datasets?



You should be able to get a successful merge. Are you working with the CEPR data available here? If so, I did download the 2007 data along with an extract of 2007 IPUMS CPS data and was able to get some matches on a one-to-one merge. You should be able to do so with the following variables, as they would uniquely identify an observation in a given year: MONTH, HRHHID, HRHHID2, and LINENO.

I did have to make some changes to the CEPR data in order to get the merge to work. Most notably, I created numeric values of the household identify variables (hhid, hhid2) in the same format as IPUMS data (Stata code below):

destring hhid, generate(hrhhid)
format hrhhid %15.0f
destring hhid2, generate(hrhhid2)
format hrhhid2 %12.0g

Additionally, it looks like some of the coding for the CEPR data is different from IPUMS, specifically in this case, sex. Make sure if you are trying to match on a variable, that you adjust them to use the same coding.

I hope this helps. If you are still running into issues with the merge, you can email with your code.