Objectives of the Study

Despite its limitations and problems, the decennial census long form sample is still regarded by most as the “gold standard” for data at the county and sub-county levels. No other survey has the sample size, scope or breadth of material that the decennial census long form possesses. These data, provided every 10 years, are representative and comprehensive. With the advent of the ACS, we now have the opportunity to obtain a comprehensive view of county and sub-county areas more than once a decade -- if the American Community Survey’s rolling sample can deliver on its promise to provide a picture at least as good as that provided by the decennial census long form.

With this in mind, the objectives of this analysis are:

1. To compare the quality of data for the two surveys, including measures of initial cooperation, overall unit nonresponse, and item-specific imputation.
2. To determine whether the socioeconomic and long form housing information on the ACS is similar to that provided by the decennial census (for roughly equivalent time points) at the county and sub-county levels. The focus is on meaningful differences between the two surveys, defined as statistically significant differences of over two percentage points at the county level. While tiny differences may be statistically significant, they often do not have a substantive impact in a “real world” setting. On the other hand, when a meaningful difference exists between a variable measured in the ACS and census, a planning study could come up with a dissimilar framework, depending on whether the variable used was from the ACS or census.
3. To examine the geographic pattern of differences by neighborhood and by other variables, including quality measures, that could shed light on the reasons for differences
4. To determine how the use of the ACS versus the decennial census sample affects real-life applications of neighborhood data related to City Planning. A case study of a City Planning-related application is used to determine whether use of the ACS instead of the census would result in substantive differences in program planning and targeting.