I just wanted to double check if the following variables from original CPS correspond to correct variables in IPUMS CPS:
incwage corresponds to wsal-val (-1987)
oincwage corresponds to ws-val (1988+)
incbus corresponds to semp-val (-1987)
oincbus corresponds to se-val (1988+)
incfarm corresponds to frse-val (-1987)
oincfarm corresponds to frm-val (1988+)
inclongj corresponds to ern-val (1988+)
The correspondences between the income variable names in the original ASEC microdata and in IPUMS CPS for different time periods are can be found in the tables on the income topcodes page.
Great, thanks! This is really helpful! I have two follow-up questions:
Since ern-val corresponds to inlongj, i-ernval should correspond to qinlong. But then I am confused why i-ernval has only two categories (0 - no change or children; 1 - allocated) and qinlong has four categories (0 - no allocation; 1 - income amount allocated; 2 - income type allocated; 3 - income amount and type allocated). I am referring to 1993 code book for i-ernval, but I think it is the same for other years.
qoinbus and qoinfarm are data quaility flags for oinbus and oinfarm, respectively. Although oinbus and oinfarm start in 1988, qoinbus and qoinfarm are available for earlier years. This means that qoinbus and qoinfarm are data quality flags for incbus and incfarm the years before 1988?
Regarding your first question, QINCLONG actually incorporates information from two separate flags, one of which (I-ERNYN) tells whether the yes/no question about having earned income was imputed, and the other (I-ERNVAL) which gives information about the value of the earned income in the longest job. QINCLONG=3 if both were imputed. I don’t know the details of how the Census Bureau did the allocation, for example how can there be cases where ERNVAL is not imputed, but ERNYN is imputed?
I realize this is not noted in the variable description, and I’ve notified the IPUMS CPS team that the documentation of this variable could be clarified.
ERNYN is available as an unharmonized variable in IPUMS CPS, UH_INCSER_A1, while I-ERNYN is available as UH_AINCSER_A1.
Regarding your second question, I am currently investigating this and should be able to report back within a few days.
@sheldon it turns out that IPUMS CPS does have an unharmonized variable for I-ERNYN. It is called UH_AINCSER_A1. I’ve modified my previous post to reflect this. If you cross-tabulate UH_AINCER1_A1 and UH_AINCSER_A1, you’ll find the distribution is exactly the same as QINCLONG.
@sheldon The flags QOINCBUS and QOINCFARM should actually be separate variables called QINCBUS and QINCFARM for the pre-1988 period. We had a similar confusion with QINCWAGE a few years back. I’ve notified the IPUMS CPS team that these variables need to be added. In the meantime, you can treat QOINCBUS as the flag for INCBUS pre-1988, and similarly for QOINCFARM. Thank you for pointing out this error. We like to show appreciation to our data users who point out errors such as this by sending them an IPUMS mug. Please email ipums@umn.edu with your address, and we’ll get you one in the mail!
great, thanks!