I would like to identify whether reported income amounts (such as inctot) were actually reported by the person being asked the question or whether the inctot data that is in the Census was given by a proxy repsondent (such as a spouse, etc.) Is there a way to determine whether people were the actual respondents to the question or whether the data reported for them were entered by a proxy respondent ?
For 1990 Census and 2000-2011 ACS samples.
For ACS data 2000-2011 samples there are data quality flag variables that indicate if income responses are given directly by the respondent or if the value has been edited or allocated in some way. These can generally be found under the “Flags” tab of the INCTOT variable, and other variables. Specifically, the data quality flag for INCTOT is QINCTOT.
Thank you for your response Jeff. I noticed that there are only observations for QINCTOT for 1950, 1960, and 2000, but I need to identify nonrespondents for 1990 (Census), and 2001-2011 (ACS). I see that there are additional flag variables that are associated with the inctot variable, but I am not sure how they relate to the inctot variable. For example, when I click on the flag tab for the inctot variable, there are over 10 flags associated with inctot (e.g. QEARN, QINC, etc.) If my goal is to obtain a sample of individuals who were only respondents for the inctot variable for 1990-2011, will I need to use these other flags variables on top of the inctot flag? If so, can you provide some guidance as to how I can do that?
Yes, if your goal is to identify individuals who were respondents for the INCTOT variable you’ll need to use the flag variables for all of the income variables that contribute to INCTOT. There is a nice discussion of the allocation proceedure on the Editing Procedure tab for INCTOT. This should provide you with the necessary guidence, but feel free to reach out with any additional questions.