2020 CENSUS: Additional Actions Could Strengthen Field Data Collection Efforts

What GAO Found

The Census Bureau (Bureau) recently completed its 2016 Census Test in Los Angeles County, California, and Harris County, Texas. One primary focus of the test was to assess the methodology for non-response follow-up (NRFU), where enumerators personally visit households that do not self-respond to the census. GAO found that during the 2016 Census Test, NRFU generally proceeded according to the Bureau's operational plan. However, data at both test sites indicate that the Bureau experienced a large number of non-interviews. Non-interviews are cases where either no data or insufficient data are collected. Bureau officials are not certain why there were so many non-interviews for the 2016 Census Test and are researching potential causes. Going forward, it will be important for the Bureau to better understand the factors that contributed to the non-interview rate because of its relationship to the cost and quality of the census.

GAO also found that refining certain enumeration procedures and training enumerators better could produce additional efficiencies by enabling the Bureau to be more responsive to situations enumerators encounter on the ground. For example, enumerators, by design, were unable to access on the mobile device recently closed, incomplete cases. Bureau officials acknowledged that closing cases in this fashion represented a missed opportunity and plan to test greater flexibilities as part of the 2018 End-to-End Test. Programming some flexibility into the mobile device—if accompanied with adequate training on how and when to use it—should permit enumerators to complete some interviews and reduce the cost of follow-up attempts. Further, enumerators did not always understand procedures for visiting property managers in multi-unit buildings. Specifically, the 2016 Census Test demonstrated that vacant units could quickly be removed from the NRFU workload where a property manager was readily available to provide that information; however, in other cases the procedures confused enumerators and they did not understand how to proceed. Without the knowledge of which units were vacant, enumerators may have unnecessarily visited some vacant units and thereby increased the cost of NRFU.

During GAO's field visits, GAO encountered several instances where enumerators learned that returning at a specific time on a later date would improve their chance of obtaining an interview from either a household respondent or a property manager. However, the Bureau's 2016 Census Test and automated case management system did not have an efficient way to leverage that information. Attempting contact at non-responding households at times respondents are expected to be available increases the completion rate and reduces the need to return.

Why GAO Did This Study

With a life-cycle cost of about $12.3 billion, the 2010 Census was the most expensive enumeration in U.S. history. To help control costs and maintain accuracy, the 2020 Census design includes new procedures and technology that have not been used extensively in earlier decennials, if at all. While these innovations show promise for a more cost-effective head count, they also introduce risks. As a result, it will be important to thoroughly test the operations planned for 2020.

The objective of this report is to assess key NRFU operations performed during the 2016 Census Test to identify any lessons learned that could have a potential impact on pending design decisions for the 2020 Census. To assess NRFU operations GAO visited both test locations and observed enumerators conducting NRFU interviews, and reviewed relevant documents including the test plan and enumerator training manuals.