Single Year or Multiyear Migration Estimates? An Analysis and Comparison Using the American Community Survey.

David Ihrke, U.S. Census Bureau
William K. Koerber, U.S. Census Bureau

Using estimates from the 2008 and 2006-2008 American Community Survey (ACS), this project analyzes in, out, and net migration flows for the population 1 year and over at the state level. By comparing the reliability of these estimates, we attempt to guide researchers in their decision over which estimate to use and when. In certain situations the robustness of the 1-year estimate is preferable, while others require the weighted average approach of 3-year. Estimates from both sources are compared to illustrate the strengths and weaknesses. 1-year estimates give researchers more timely data, but the coefficient of variance may be too large for useful analysis of smaller estimates. 3-year estimates give researchers less variability but might mask significant yearly changes. The combined use of the two datasets can provide a much more complete story of the migration between states of various population groups.