2010 Census: Census Bureau Needs Procedures for Estimating the Response Rate and Selecting for Testing Methods to Increase Response Rate

In our 2008 report we found that the response rate for the 2010 Census was not supported by detailed documentation. For the 2020 Census the Bureau has taken steps to document the components of the response rate. For example, the initial response rate estimate for 2020 incorporated 2010 Census data, including response rates by local census office and by type of enumeration area. The response rate was then used as one of the factors the Bureau modeled for the sensitivity analysis that was used to produce the initial ROM lifecycle cost estimates for the 2020 Census.

In 2008, we reported that the Bureau did not have procedures for reevaluating and updating the 2010 response rate estimate, including identifying events or changes in related operations that should trigger a review and documenting the results of such reviews. According to the Bureau, starting in 2012 they began conducting research related to alternate response modes and to the use of administrative records to substitute for personal visit nonresponse followup; and that the results of that research and testing will inform its analysis and estimation of likely response rates and nonresponse workloads for the 2020 Census. Following that research, the Bureau will conduct specific research related to estimate 2020 response rates. In addition, response rates to the American Community Survey are being monitored throughout the decade to determine if changes in response rates are occurring. Finally, the Bureau will establish a team to monitor socio-political events that may serve as triggers for a reevaluation of the response rate estimate. According to the Bureau, if an event were to occur during intercensal testing that did warrant an examination of the response rate, then the team described above would make that determination. If the response rate were revised, the decision would be documented through a formal memo.

In our 2008 review of the Census Bureau's (Bureau) testing of methods to increase response in the 2010 Census, we found the Bureau could not specify how they selected methods to test for increasing responses; and recommended the Bureau establish procedures for selecting methods for increasing response rate that will be the subject of research and testing, including requirements for documenting how the Bureau defines and weighs factors used to select methods and documentation on methods considered but not tested. The Bureau has taken action and from 2009 to 2011 established and implemented a rigorous process for selecting research projects to test and not to test. According to Bureau documentation the Bureau cataloged over 600 ideas and has now identified 26 research projects to inform the 2020 Census design.