648 – Targeted Address Canvassing and Coverage
Designing an Adaptable Database for Model-Based Research
Nancy Johnson
U.S. Census Bureau
Mary Pritts
U.S. Census Bureau
In preparation for the 2020 Census, the U.S. Census Bureau is conducting research on building statistical, model-based simulations for a selective field canvassing operation and for producing estimates of coverage error. This type of research requires a flexible and adaptable database structure that incorporates data from various sources, allowing for different levels of analysis, across different time periods, in an ad hoc manner. The starting point for the database is the universe of over 155 million records from the 2010 Census Address Canvassing operation. It then incorporates data from the final 2010 Census and various Administrative Record sources covering different spans of time. The result is a multi-use, multi-source data set fully integrated at the level of analysis most appropriate for the research project.