The ultimate success of the American Community Survey (ACS) will rest on its capacity to provide users with reliable data for small areas more than once a decade. Given the complex demography of its population and its abundance of hard-to-enumerate groups, the Bronx, NY test site poses formidable challenges for the ACS. Differences in overall response and item imputation by mode of data collection (mail versus enumerator response) highlight major differences between the ACS and the decennial census. An understanding of these differences is essential as data users attempt to assess the ramifications of eliminating the 2010 census long-form and replacing it with a continuous ACS. Furthermore, this research attempts to inform future assessments of the ACS, as operational data from the 2000 Census become available for small areas throughout the nation.