Status:Closed    Asked:Apr 13, 2017 - 10:56 AM

Person Weight

I am trying to get an accurate count of employed persons' income. I noted some incomes are 0 or similarly small. Does that mean they reported making 0 or 10 dollars in a given year? Also, the PERWT, does that mean that to get an accurate count, I would create separate line that many times for that value? Example, income 500 with a weight of 6, I would create a total of six lines with an income of 500?

Do you have the same question? Follow this Question

Staff Answer


Jeff Bloem


Yes, if INCWAGE codes are 0 this indicates that the respondent is reporting an earned wage or salary income of zero for the previous year. Regarding the sample weight PERWT see this documentation and the IPUMS FAQ for more information about how to use these weights. You intuition is on the right track, but statistical software packages, such as STATA, SPSS, and R, make using sample weights much easier than performing this calculation in excel.


Apr 13, 2017 - 12:36 PM

Report it


You have two questions here, about about measurement and one about weights. As far as the second one goes, you have to use software that supports estimation with complex survey data (Stata with svy commands, SAS with PROC SVY whatever, R with library(survey) ). Don't try this in Excel. If you don't understand what these weights are, you are probably several months away from doing this analysis properly, as you would have to learn at least the basics of underlying statistics.


Apr 13, 2017 - 11:15 AM

Report it


When did IPUMS-CPS start like the way (easily downloadable way) like now?
Replicate weights for use with CPS-TUS
Longitudinal cohort in TUS 2014/2015?
When can we expect NFHS 4 data to included in IPUMS-DHS database?
trying again need more characters
DOINGLW2 variable
Login   |   Register

Recently Active Members

View More »

Share |