The ACS Group Quarters (GQ) sample was designed to produce state-level estimates of characteristics of the GQ population. However, GQ residents are a component of the total resident population, and characteristics of the total population are published for smaller areas. There is an insufficient GQ sample to produce estimates for these smaller areas. Data users had noticed large year-to-year fluctuations in estimates of counties, and a lack of representation of GQ sample across small areas had been documented. In order to address these concerns, the GQ population sample has been supplemented by a large-scale, whole person imputation into not-in-sample GQ facilities starting with the 2011 ACS data release, which includes the 2011 ACS 1-year, 2009-2011 ACS 3-year, and 2007-2011 ACS 5-year estimates. Roughly as many GQ persons are imputed as interviewed.
Estimates for substate geographies of the total GQ population and of characteristics of the total resident population will be more reliable than earlier years with the previous methodology. The state-level estimates of characteristics of the GQ population will be relatively unchanged by the new methodology.
Data users should notice little or no difference regarding ACS data products. In particular, the Census Bureau will continue to produce only state-level estimates of characteristics of the GQ population. We do not plan to release estimates of characteristics for substate areas for this population in the future. There is a break in series between the 2010 and earlier ACS data products and the 2011 ACS data products. Namely, the 2010 and earlier ACS data products only have sample GQ person data, whereas the 2011 ACS data products have both the imputed and sample GQ person data for years 2007-2011. This is particularly relevant for substate areas where the GQ population is a significant portion of the total resident population. For more information, read the full ACS Group Quarters Small Area Estimation user note.