The research and advocacy community has now come to accept the fact that any remedy for the undercount of young children in the Census must go beyond increased education and outreach efforts, as best exemplified by a recent Blogpost on this website. The focus on outreach, while necessary, has proven insufficient to address this seemingly intractable problem. It has been said, however, that every problem — even one as difficult as the undercount of young children — can give rise to opportunities for innovation.
One such opportunity to address the undercount of young children is related to the creative use of administrative records to improve household rosters. Such an approach is consistent with Enterprise efforts at the Census Bureau, to reconceptualize the census enumeration as encompassing data collection from multiple sources.
The epitome of these data collection efforts is the Early Childhood Integrated Data Systems (ECIDS). ECIDS exist in more than 30 states, and are very important in the formation of policies and the execution of strategies to address need. ECIDS provide information on young children from a variety of administrative sources, which include data on persons receiving various types of benefits. Substantial research, which is summarized here, has shown that undercounts are more likely to be borne by Black and Hispanic children, the very children who are over-represented in needs-based state and local administrative records. Most important, these data are compiled by persons with substantial on-the-ground knowledge of what the data represent, the very knowledge that the Bureau needs in order to assess their utility. In other words, these data are put to use in an active way by governments. (More information on ECIDS may be found here.)
These administrative data can be used to enhance what has come to be known as the Census Bureau’s Demographic Frame, a database of person-level information that includes demographic characteristics and addresses for individuals. Among other things, it’s used by the Census Bureau to improve data quality, with an eye for use in the improvement of questionnaire rosters in the 2030 Census. The Demographic Frame is derived from a variety of data sources, which include administrative data and – most important – provides unique identifiers for each person that allow individuals to be linked across different datasets, including those from the decennial census.
A Use Case on the undercount of young children developed by staff at the University of Virginia has explored the prevalence and content of ECIDS at the state level and identified a number of states that have well-developed systems, with detailed data and documentation. This Use Case proposes an experiment on the capacity of ECIDS to enhance census rosters with information on
young children, using the results of the 2020 Census for selected states. It is time to consider a 2020 Census pilot study, where agreements are established initially with one or two states, to determine whether the individual records contained in ECIDS can enhance the demographic frame and provide information on young children beyond what is currently available from other sources. While it is unlikely that any ECIDS – by itself — can provide a comprehensive count of all children in a state, the key question is whether ECIDS records can help the Census Bureau improve coverage of young children by reducing their undercount by Race and Hispanic origin. As a way of assuring state governments that use of these data is restricted to applications aimed at improving the coverage of young children, agreements with participating states need to be well-defined in plain language.
Having been made a priority by the Census Bureau, the undercount of young children begs for innovation, the kind that ECIDS may bring to the table. Those who advocate for better counts of young children need to encourage the Census Bureau to explore ECIDS as a source of administrative records and support the provision of resources to make it happen.
