Calculating MOEs from derived ACS estimates

Over the years, we have heard from a lot of people interested in getting an easy-to-use tool that would help people calculate margins of error from derived ACS data (e.g., data combined across categories or geographies). There are several organizations that have developed some basic applications that might be useful. Here are the links:

sdcclearinghouse.wordpress.com/.../
www.psc.isr.umich.edu/.../
pad.human.cornell.edu/.../
fyi.uwex.edu/.../
www.demography.state.mn.us/.../StatisticalCalculationsMenu.xls

If you are using a different application in your organization, feel free to post it here.
  • Isn't it long past time for the Bureau to take care of this problem in AFF, providing MOEs for aggregated data? It was promised at the beginning of AFF2 development, but never came to pass.
  • We calculate MOE as described in the Compass series as mentioned in the first post. When we aggregate across multiple categories, if there are multiple 0 estimates, then we count the MOE for the 0 estimate only once. That reduces the aggregated MOE somewhat, but even so, we often get MOEs of over 100% of the estimate in some small geographies.
  • Thanks for the replies.

    My problem and the answers I got from the Census this week is exactly the same as what Gene described last October. So I guess there is still no solution. It is so disappointing to know that I cannot use the data in the way I wanted.
  • In reply to Lin Song:

    Old thread but placing here in case others stumble across it like i did. The census has a very good section with several examples of the pitfalls or aggregating MOE. See pg 25 in American Community Survey Multiyear Accuracy of the Data (5-year 2012-2016) www2.census.gov/.../MultiyearACSAccuracyofData2016.pdf
  • Old thread but placing here in case others stumble across it like i did. The census has a very good section with several examples of the pitfalls for aggregating MOE. See pg 25 in American Community Survey Multiyear Accuracy of the Data (5-year 2012-2016) www2.census.gov/.../MultiyearACSAccuracyofData2016.pdf