ACS Data Users Group

I'm trying to replicate how the ACS calculates their MOEs for Gross Migration and NetMigration from the latest 2013-2017 County-to-County Migration flow data. I followed their guidance/methodology for approximating a MOE for derived estimates but got different results from what was published in their data table. Does anyone know how to calculate a MOE for Net Migration and Gross Migration?

When you say that your estimates are "different," how different are they? The formula for derived MOE does not match published MOEs exactly (and gets farther from the published MOE as more individual elements are combined in the derivation).

Thanks Beth for replying!!

The Census Bureau published 2013-2017 County-to-County migration flows gross migration for Travis County, Texas(Geography A) to Hays County, Texas (Geography B) as: Estimate-11,214 and MOE-1,369. Following their methodology to get a derived estimate for gross migration, I get the same estimate but MOE is 1,438. The difference is small but there is still a difference. None of the MOEs for gross migration match when using the methodology published by the Census. There is no guidance on how to calculate Net Migration MOE, which is different than gross migration though it uses the same estimates and MOEs from Geography A and Geography B.

Do you have any guidance on how to calculate Net Migration MOEs? And, is the difference found in the published MOE from the MOE calculated using guided methodology meaningful?

I'm not even sure where to begin on the difference between published and derived MOEs, except to say that this is an issue data users have been struggling with since the ACS was first launched. There is no satisfying answer--derived MOEs will differ from those published. The differences are a known problem, but one that has not yet been resolved (though for some products there are now better tools for calculating MOE, such as www.census.gov/.../variance-tables.html).

For calculating MOE for net migration (aside from what's published), those are tables I don't work with on a regular basis, so I'm hoping someone else will chime in on this thread.