Status:Closed    Asked:Apr 24, 2018 - 10:37 PM

Is there some alternative to quality flags that gives information about imputation of income variables?

Neither HHINCOME nor FTOTVAL nor INCTOT have qualuty flags under the flags tab, even though we know that some values come from actual answers and others from hot-deck imputation or similar proceedures. This seems odd, as these are the variables where I would expect quality to be highly important. Research has indicated that imputed values are not good substitutes for actual values. See, e.g., "Trouble in the Tails? What We Know about Earnings Nonresponse Thirty Years after Lillard, Smith, and Welch"

Is there anything you can tell me about this? Is there, for example, a quality flag for all the income values collectively? Or do the replication weights implicitly incorporate imputation by increasing the relevent varience? Are there quality flags for income in the Basic Monthly or Suppliment studies that would provide some, perhaps indirect, information about the quality of income in the ASEC? Or is there some other information in the ASEC or elswhere in the CPS that someone wishing to assess the quality of the income data should know about?

Do you have the same question? Follow this Question

Staff Answer


Jeff Bloem


This is a great question and one that I think could be explained in better detail. We will work on making this improvement, but in the meantime here are the details.

Income variables such as HHINCOME, INCTOT, and FTOTVAL are all income variables that combine values of other more specific income variables. For example, the comparability tab for INCTOT describes the components of INCTOT in various sample years. In ASEC samples from 1988 and onward, INCTOT includes values from INCWAGE, INCBUS, INCFARM, INCSS, INCWELFR, INCRETIR, INCSSI, INCINT, INCUNEMP, INCVET, INCSURV, INCDISAB, INCDIVID, INCRENT, INCEDUC, INCCHILD, INCALIM, INCASIST, INCOTHER.

Some of these variables have their own data quality flags. For example: INCASIST has QINCASSI and INCCHILD has QINCCHIL. Others, like INCWAGE and INCBUS, are themselves aggregated values of multiple variables. For example: INCWAGE is derived from OINCWAGE and INCLONGJ which both have data quality flags, QOINCWAGE and QINCGONG respectively.


Apr 25, 2018 - 08:56 AM

Report it


May I suggest an easy way to do this?

Add a "Qualituy Flags" column to the comperability tabs for HHINCOME, INCTOT, and FTOTVAL showing a "*" for the variables with quality flags, and a note to that effect in the "Flags" tab. If you wanted to be extra-helpful, you could place a similar note in the flags tabs of other variables that are income subtotals, have no flags, and have components with flags, but I don't think that is essential.


Apr 26, 2018 - 06:33 PM

Report it


Resources on imputation of income in the IPUMS?
1.With data quality flags, do file won't finish, "qage already defined." 2.On...
Census samples: sampling design and data quality
Is there a list of IPUMS-CPS ultimate income components in somewhere? Of inco...
Login   |   Register

Recently Active Members

View More »

Share |