Status:Closed    Asked:Oct 07, 2013 - 10:04 AM

Are there "control totals" for the March 2012 CPS to insure I'm getting correct counts of total persons and hholds

I've generated tabulations of persons and households by hhold type but the results seem a little high. I'm tabulating only household heads (relate = '0101') and I get 121 million households. Other sources show the count closer to 117 million and with the Census websites shutdown I can't get to any CPS reports to verify my numbers. I'm wondering if there are any established control totals that can be used to verify that tabulations result in the correct figures.

Do you have the same question? Follow this Question

Staff Answer

Unfortunately, I'm not aware of any official published figures outside of those that would typically be available through a government website. Off the top of my head, the only explanation for the discrepancy that I can think of would be that the other totals you are looking at use a different weight, such as A-FNLWGT. This variable isn't available through IPUMS but is in the public-use file available through NBER. Though I'd be surprised if the weight made that much of a difference, I suppose it is at least possible.


Oct 07, 2013 - 11:12 AM

Report it


Does anyone have sample code for using svydesign function in R?
Which variables should I use in my data extract for determining father-son pa...
What "Psychiatric, Nursing, and Home health aid" occ 3600 encompasses?
Difference in Sample Size between ACS 2016 and Census Technical Documentation...
Incomplete alphabetic variable list and broken IPUMS group table in thee CPS.
What R package is needed for IPUMS USA for weighting? Is it "survey"? Do you ...
Login   |   Register

Recently Active Members

View More »

Share |