The American Community Survey (ACS) includes a "year-of-entry" question that asks the foreign born when they came "to live in the United States." Researchers have used these data to analyze "recent" immigrants -- usually defined as arriving within the last five to 10 years -- noting that the number of new arrivals is considerably lower than those who arrived in prior years and suggesting the size of the recent immigrant population may be underestimated. This paper shows the primary cause of the lower-than-expected estimates to be the ACS data collection methodology, which causes an inherent downward bias in survey-year arrivals. [ABSTRACT FROM AUTHOR]