The purpose of this national level study is to examine the cost reduction and coverage impact that would result from conducting a targeted Address Canvassing operation. This study was conducted with 2009 vintage data, to simulate a 2010 Targeted Address Canvassing operation. As examined here, TAC is defined and studied as a way to identify geographic areas (in this analysis, Census 2000 current blocks) using statistical models to identify the most cost beneficial updates to the Master Address File. Statistical modeling is one approach to targeting blocks for Address Canvassing that may result in cost savings. The research presented here indicates that there is substantial potential for cost reduction using a model-based TAC approach, and make six recommendations based on our experiences conducting this research.
A nationwide Address Canvassing operation was used in 2009 to update the Master Address File in preparation for the 2010 Census. Other updating procedures involved using Administrative Record data from the United States Postal Service in the form of the Delivery Sequence File and address files from the Local Update of Census Addresses program. The workload for the 2010 Address Canvassing operation totaled over 150 million address records (mostly Living Quarters). While this was a valuable endeavor, it was also very expensive; with the Census Bureau incurring approximately 459 million dollars in execution costs (lister training, lister salary and benefits, lister mileage) and about an additional 400 million dollars in other costs (materials/equipment, infrastructure and contract costs, etc.).