Status:Closed    Asked:May 07, 2015 - 02:37 PM

Does the IPUMS team simply divide the weight by three when releasing the three year file of the ACS?

I'm wondering if I should concatenate the data myself or if should wait to the latest release of ACS three year data file. I would like my data to be as close with the IPUMS release and with the census. If the team does more than simply dividing the weihght by three I'd prefer to wait. Please let me know/

In a previous response you mentioned that the data was likely to be released on April 2015. Is the latest three year file going to be release sometime soon??

Do you have the same question? Follow this Question

Staff Answer




While the Census Bureau does use more complex methods for generating multi-year weights (see Weighting Methodology from the 2012 3-year Accuracy Statement ), simply dividing by three is the primary adjustment and provides extremely similar values to the Census Bureau calculated weights. You can verify this by trying to match the 2011-2013 PUMS Estimates for User Verification. See this answer for more details on constructing the multi-year files yourself.

The Census Bureau has proposed discontinuing the 3-year ACS files. IPUMS will continue to provide the 3-year files as long as the Census Bureau releases them. Generally IPUMS-USA releases multi-year ACS samples within 2 months of their initial release by the Census Bureau. This year the IPUMS-USA team is working on improving their data creation process. This will ultimately improve the IPUMS-USA project as a whole, but with the unfortunate side-effect of delaying the release of the 2013 multi-year ACS data files. At this time, there is no estimated release date for the multi-year samples.

Hope this helps.


May 07, 2015 - 04:10 PM

Report it


Trying to do ACS weights yourself be quite silly. Census Bureau develops weights for their 3-year and 5-year ACS data products, and IPUMS data simply reproduces these weights. (Note that the 3-year ACS is discontinued for budget reasons.) The 3-year weights correspond to some imaginary, never-existed population spanning years 2011-13, say; while each of the single-year weights would correspond to their (never-existed, as well, but arguably better defined) population in that year. These populations differ because of births, deaths, and migration. Census statisticians adjust the weights in each of the data products to agree on multiple dimensions internally and with extrnal demographic figures (population projections), and you simply don't have the resources to reasonably reproduce their work.



May 07, 2015 - 03:00 PM

Report it

oh, ok. I read someon from the IPUMS team suggesting to simply dived the wights in this forum...

Just to be clear, is IPUMS discontinuing the production of three year file?? or the census?


May 07, 2015 - 03:05 PM

Report it


Census is releasing the 2012 ACS 1-year file on Dec 17 -- When will it be ava...
When will 2015 ACS data be included in IPUMS-USA?
What was the actual release date of the IPUMS 2015 ACS 1 year data set?
Timing for 2016 ACS 1-year release?
Login   |   Register

Recently Active Members

View More »

Share |