Introduction

The Census Bureau began full implementation of the American Community Survey (ACS) in 2005. The ACS will replace the function of the decennial long form, obtaining comparably detailed information released annually. While the long form represented a snapshot in time, April 1, 2000, the ACS interviews monthly samples (U.S. Census Bureau, 2006a) to produce three period estimates: one based on one year of collected data, one based on three years, and one based on five years. The purpose of this paper is to highlight key issues of interpretation for data users. It contributes to the development of guidelines which will aid ACS data users in interpreting the ACS one-year, three-year, and five-year estimates, and in choosing which of these estimates is most appropriate for their data needs. The users we are targeting include both statistically unsophisticated users and sophisticated users who could benefit from an introduction to these new data products. This paper presents results of work in progress.

Summary: When and When not to Use Multi-Year Estimates

When to Use Multi-Year Estimates

• For tracts and other smaller geographies - single-year estimates are not available.
• To obtain estimates with lower standard errors.
• For smaller subpopulations of larger geographies.
• For more precise comparisons of change over time (non-overlapping comparisons).
• For smoothing data over time.

When to Use Single-Year Estimates

• For larger geographies and populations - prefer more current single-year estimates.
• For counties and states for age, sex, race, Hispanic origin, or housing units, where the standard errors are already small due to controlling.