In the early 1990s, the Census Bureau proposed a program of continuous measurement as a possible alternative to the gathering of detailed social, economic, and housing data from a sample of the U.S. population as part of the decennial census. Pilot testing of the new survey began a few years later, and a major expansion of the survey’s sample was undertaken as a formal experiment in the 2000 census. The replacement of the “long-form sample” with the new, ongoing survey—and the consequent casting of the 2010 decennial census as “short form only”—became a key part of the Census Bureau’s strategy for the 2010 census. The full-fledged American Community Survey (ACS) became a reality in 2005 with full-scale, nationwide implementation at the household level.
The design of the ACS relies on monthly samples that cumulate to approximately 10 million completed interviews over each 5-year rolling interval. The data products are period—as opposed to “point”—estimates based on averaging 1, 3, or 5 years of data. Beginning in 2006, the Census Bureau has published annual 1-year estimates of characteristics of the U.S. population and housing units for all geographic entities with populations of at least 65,000. Since 2008, 3-year estimates for geographic entities with populations of at least 20,000 have been also been reported. The end of 2010 will mark a crucial milestone for the ACS. In December 2010, the first set of estimates based on 5 years of continuous data collection will be published for all statistical, legal, and administrative entities, including areas as small as census block groups.
When the ACS entered full-scale production in 2005, it did so only for the household population. ACS coverage of what the census refers to as group quarters (GQ) —places such as college dormitories, nursing facilities, in-patient hospice facilities and military barracks—began one year later, in 2006, primarily to more closely replicate the design of the census long-form sample.
Box 1-1 provides the definition of group quarters used by the Census Bureau for purposes of the ACS and lists the major group quarter types included in the survey.
For the exact wording of the panel’s charge, see Appendix A.
Group Quarters Definition and Major Types
A group quarters is a place where people live or stay, in a group living arrangement, that is owned or managed by an entity or organization providing housing and/or services for the residents. This is not a typical household-type living arrangement. These services may include custodial or medical care as well as other types of assistance, and residency is commonly restricted to those receiving these services. People living in group quarters are usually not related to each other. Group quarters include such places as college residence halls, residential treatment centers, skilled nursing facilities, group homes, military barracks, correctional facilities, and workers’ dormitories.
SOURCE: U.S. Census Bureau (2008a).
In 2007, data products based on the ACS began to phase in sample-based estimates of the characteristics of the total (household and GQ) population. The 2010 release of 5-year period estimates will be based on one year of data (2005) that exclude a GQ sample and four years of data (2006-2009) that include GQ samples. Beginning in late 2011, all ACS data products will be based on samples of both households and group quarters.
The 5-year period estimates released in late 2010 from the ACS will furnish population characteristics for the same statistical, legal, and administrative entities as the 2000 census long-form sample. This includes entities as small as census block groups. Although ACS data products are expected to evolve based on data needs and feedback from researchers, the products that will be a part of this first 5-year release will be similar to existing ACS products. Selected data tables will report a breakdown of the total population into those living in households (often accompanied by additional characteristics) and those living in group quarters (with no additional detail).
In some small geographic areas, GQ residents represent a large proportion of the total population, whereas in other areas their number is relatively small or zero. The
accurate counting of the GQ population in the decennial census has been an ongoing concern (National Research Council, 2004), and the measurement of this small (relative to the whole population) but diverse component of the population poses new challenges for the ACS.
In early 2010, the Census Bureau asked the Committee on National Statistics of the National Academies to convene a panel to examine and comment on the Bureau’s statistical methods for measuring the GQ population in the ACS. The panel was asked to recommend alternatives that can make the data more useful for small-area data users, particularly users of ACS 5-year period estimates for small governmental entities, considering operational feasibility and compatibility with the treatment of the household population in the ACS.
This interim report makes recommendations addressing near-term improvements in the sample design, weighting, and estimation of the GQ population. At the end of the 24-month study, the panel will issue a final report with conclusions and recommendations for longer term improvements in light of data user needs.
The replacement of the census long-form sample with the ACS promises data users major benefits as well as new challenges. In terms of benefits, the critical advantage of the ongoing, continuous ACS is the timeliness of the estimates and the increased frequency of data releases. The continuous measurement also has some advantages in terms of data quality. Whereas the decennial census relies heavily on a vast temporary workforce that must be hired, trained, and deployed quickly, the continuous nature of the ACS can accommodate a staff of well-trained, permanent field representatives. This, in turn, may contribute to reducing various kinds of nonsampling error, including item nonresponse rates compared with data from the 2000 census long-form sample.
However, the ACS has some offsetting disadvantages, such as higher sampling errors associated with the estimates—a consequence of the smaller overall sample size (compared with the 2000 census long-form sample), even cumulating over 5 years, and the strategy of sampling for nonresponse in the follow-up stages of ACS data collection. In addition, the control totals used to reduce variation in the ACS estimates are based on postcensal population estimates, instead of the census itself, which means that the errors associated with the population estimates will also affect the ACS estimates.
The decision to include group quarters in the ACS presents some new challenges and data quality implications. One of these challenges stems directly from the survey design. Currently, a stratified sample of group quarters is selected for each state, without controlling for the allocation of GQ populations at substate levels of geography, such as counties, municipalities, tracts, and block groups (unlike the 2000 census long-form sample, which was generally controlled to census counts at subcounty levels of geography). Consequently, the measurement and estimation approaches developed for the GQ population are designed to be optimal only for estimates at the state level and higher levels of geography—yet the household sample is suitable for producing estimates of characteristics of people residing in households for substate geographies.
Another challenge posed by the inclusion of group quarters in the ACS is “sampling zeroes”—counties and smaller geographic areas that have no group quarters represented in the sample, even after a 5-year period of data collection, despite the presence of GQ populations in those areas (including in the most recent decennial census counts for those areas). This adversely affects the estimates of characteristics for the total population, as well as the usefulness of the data about the GQ population itself, and it can elevate estimated standard errors for characteristics of the total population. The timing of the first 5-year data release for small areas—followed in a few months by the release of 2010 census counts—will make this especially awkward when the census counts reveal a nonzero GQ population for these same areas.
Other challenges include the complexities and costs associated with maintaining an accurate and up-to-date inventory of GQ facilities, independent of the inventory of household addresses. This is especially difficult in the case of smaller group quarters, which appear and disappear faster than larger facilities, and group quarters in structures that may have been recently converted from housing units. This affects the efficiency of the GQ sampling frame, as well as the GQ estimates produced by the Census Bureau’s Population Estimates Program (PEP), which are used as controls in the ACS.
An important aspect of the panel’s charge that is not fully explored in this report—but will be a major focus of the final report—is the importance of the ACS GQ estimates for the wide community of ACS data users. A key question is whether there is a demonstrated need for GQ data in the ACS and an evaluation of the benefits of such data relative to the costs of collecting them. At present, the principal justification for including a sample of the GQ population as part of the ACS appears to be based primarily on the original vision for the ACS to serve as a replacement for the census long-form sample. The long-form sample included both institutional and noninstitutional group quarters, and GQ facilities are currently included in the ACS, at least in part, to remain faithful to that goal.
The lack of clarity regarding the need for direct estimates of the GQ population in the ACS complicates a cost-benefit analysis of this data collection effort. At this stage in the evaluation, suffice it to say that due, in part, to the complex nature of the data collection procedures applicable to group quarters (see Box 1-2), the costs associated with the GQ data are higher than those of the household data. A cost of approximately $89 per interview among the GQ population contrasts with a per-person cost of roughly $13 among the household population (a factor of almost seven); overall the GQ data collection represents around 7 percent of the ACS budget request for fiscal year 2011.
As part of its work, the panel will assist the Census Bureau by assessing data user needs and ultimately will make recommendations regarding ACS methodology in light of these needs. Issues such as whether the ACS is a suitable vehicle for the measurement of the GQ population, the data collection costs, and the importance of the data to users will be examined in the panel’s final report.
Group Quarters Data Collection Steps
The Census Bureau’s National Processing Center (NPC) mails an advance letter and brochure about the ACS to each sampled GQ facility prior to the beginning of the field work. Field representatives (FRs) contact sampled group quarters by phone to schedule an appointment for visiting the facility, and the Regional Office assists the FRs, as needed, in gaining cooperation. During the visit to sampled facilities, FRs administer the computer-assisted Group Quarters Facility Questionnaire (GQFQ) to a contact. The facility type, population size, and the sample of individuals to be interviewed are determined during this process.
Person-level interviews can be completed by:
If a GQ contact is involved in distributing the questionnaires or providing responses, he or she must take an oath of nondisclosure, under Title 13 of the United State Code.
In remote Alaska, the GQ data collection is conducted twice a year, from January through mid-April and from September through mid-January.
Correctional and military facilities selected into the sample multiple times are visited only once, during a randomly selected month.
Data collection in federal prisons is completed during a 4-month period, from September through December. The Bureau of Prisons provides a list of inmates to the Census Bureau and conducts security clearances of FRs who will be visiting these facilities.
SOURCE: U.S. Census Bureau (2009).
OVERVIEW OF THE INTERIM REPORT
This report makes recommendations that address some immediate needs in light of the data release schedule, which calls for the publication of the first set of 5-year period estimates before the end of 2010, and publication of updated estimates every year thereafter. While the panel’s initial deliberations and recommendations cannot affect the estimates in this first release of 5-year data, the intent is to use an evaluation of the initial release to potentially improve the ACS estimates planned for release in late 2011. Specifically, this report focuses on statistical methodology and proposes short-term revisions as well as research needed to inform longer term methodological decisions. Chapter 2 discusses the ACS sampling frame development and scope of coverage for the GQ component of the sample. Chapter 3 focuses on sample selection and allocation, and Chapter 4 discusses weighting and estimation procedures. The final chapter outlines the panel’s plans for the remainder of the study and proposed areas of inquiry for the final report.