Status:Closed    Asked:Sep 21, 2015 - 12:21 PM

Is there any way to use 1-year ACS estimates to approximate the discontinued 3-year estimates?

I know the weights are different, but I don't really understand how, exactly, they differ. Is there any way, for example, to adjust the weights to try to combine 1-year files into something comparable to the 3-year estimates? What are likely to be the problems with such an approach? Are the replicate weights less susceptible to these problems than the household or person weights?

Do you have the same question? Follow this Question

Staff Answer




It is possible to create your own 3-year estimates for the ACS. The multi-year files are simply a concatenation of the single-year ACS data files, which means you can select the three years you are interested in from the IPUMS-USA selection menu, adjust the dollar values to the most recent year in the 3-year set (using YEAR and the CPI99 variables), and divide all weight values by three. 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.

To test this combination method, I would recommend replicating an analysis using a three year period of the ACS that already has a 3-year summary file, such as 2010-2012. You can construct your own 3-year file and then compare your results to those already published to assess where there may be differences.

I hope this helps.


Sep 22, 2015 - 11:01 AM

Report it


How are weights for a single year sample modified to weights for multi-year f...
Should ACS person weights sum exactly (within state) to census state populati...
How precise are household counts in the City variable for the 1 year ACS esti...
Calculating replicate weights when creating 5YR files using 1YR estimates
Login   |   Register

Recently Active Members

View More »

Share |