1
Introduction

For decades, policy makers, planners, researchers, the media, and the public have looked to the decennial census as the source of detailed information on the numbers and characteristics of the U.S. population for the nation, states, metropolitan areas, counties, cities, towns, school districts, and neighborhoods. The census provides complete counts of people by such basic characteristics as age, sex, race, and ethnicity for areas as small as city blocks. It has also provided estimates from a very large sample—called the long-form sample—for areas as small as groups of blocks on people’s education, employment, income, disability, commuting, and other characteristics and about the housing in which they live. Other household surveys, such as the Current Population Survey, the Survey of Income and Program Participation, the American Housing Survey, and the National Health Interview Survey, provide more frequent, detailed information on a variety of topics, but estimates from these surveys are generally available only for the nation as a whole or, sometimes, for states or large metropolitan areas.

In late summer 2006, the U.S. Census Bureau released 2005 data from a major new continuous survey designed to provide small-area data, the American Community Survey (ACS). Each month, the ACS questionnaire—similar in content to the census long form—is mailed to 250,000 housing units across the nation; as with the long-form sample, response to the ACS is required by law. Two big differences from the long-form sample are that the ACS is conducted on a continuous basis instead of once every 10 years and the data are released every year. Over the course of time, the ACS will provide detailed data for all of the small areas covered by the long-form



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 13
Using the American Community Survey: Benefits and Challenges 1 Introduction For decades, policy makers, planners, researchers, the media, and the public have looked to the decennial census as the source of detailed information on the numbers and characteristics of the U.S. population for the nation, states, metropolitan areas, counties, cities, towns, school districts, and neighborhoods. The census provides complete counts of people by such basic characteristics as age, sex, race, and ethnicity for areas as small as city blocks. It has also provided estimates from a very large sample—called the long-form sample—for areas as small as groups of blocks on people’s education, employment, income, disability, commuting, and other characteristics and about the housing in which they live. Other household surveys, such as the Current Population Survey, the Survey of Income and Program Participation, the American Housing Survey, and the National Health Interview Survey, provide more frequent, detailed information on a variety of topics, but estimates from these surveys are generally available only for the nation as a whole or, sometimes, for states or large metropolitan areas. In late summer 2006, the U.S. Census Bureau released 2005 data from a major new continuous survey designed to provide small-area data, the American Community Survey (ACS). Each month, the ACS questionnaire—similar in content to the census long form—is mailed to 250,000 housing units across the nation; as with the long-form sample, response to the ACS is required by law. Two big differences from the long-form sample are that the ACS is conducted on a continuous basis instead of once every 10 years and the data are released every year. Over the course of time, the ACS will provide detailed data for all of the small areas covered by the long-form

OCR for page 13
Using the American Community Survey: Benefits and Challenges sample. With the advent of the ACS, the long-form sample will be dropped from the 2010 and successive censuses; consequently, the decennial census will now include only a short form with basic questions on age, race, sex, ethnicity, household relationship, and housing tenure (owner or renter). Each summer and fall from 2006 forward, the Census Bureau will release ACS statistics for the previous calendar year for areas with 65,000 or more people. In addition, by 2008, enough responses will have accumulated over the 3-year period 2005–2007 for the Census Bureau to release statistics in the fall for areas with at least 20,000 people. Finally, by 2010, enough responses will have accumulated over the 5-year period 2005–2009 for the Census Bureau to release statistics in the fall for all areas, including very small places and neighborhoods. Each year, the 1-year, 3-year, and 5-year estimates will be updated to reflect newer data.1 The implementation of the ACS represents a seismic shift in the landscape of small-area data on the U.S. population. This shift promises important benefits to users in terms of much more timely and up-to-date information than the long-form sample could provide. However, there will inevitably be a learning curve and costs in time and other resources of users to make the transition from the once-a-decade long-form sample to the continuous ACS. 1-A PANEL CHARGE Recognizing the need to assist users in the transition from the long-form sample to the ACS, in 2004 the Census Bureau asked the Committee on National Statistics of the National Academies to convene a Panel on the Functionality and Usability of Data from the American Community Survey. The charge to the panel was to study the effects of using small-area ACS estimates that are based on multiyear measurements released every year for applications that previously used static, one-time estimates from the long-form sample. The major goals of the panel’s work are to provide a base of information to ease the transition from the long-form sample to the ACS for a wide variety of data uses and to explore methodological issues raised by the use of this survey. The panel undertook a range of activities to respond to this charge. We listened to groups of small-area data users on several occasions, including at meetings we organized with major federal agency users and experienced state and local government users and at a special session of the October 2004 conference of the Association of Public Data Users. The panel also commissioned papers on the properties of different types of multiyear 1 Similar data products will be available for areas in Puerto Rico from the Puerto Rico Community Survey (see Box 2-1).

OCR for page 13
Using the American Community Survey: Benefits and Challenges estimates and the effects of using census-based population estimates as controls to smooth the ACS estimates. Subgroups of panel members met with Census Bureau staff to learn about as many aspects as possible of the ACS data collection and estimation process, including the construction of the Master Address File (MAF), from which ACS sample addresses are drawn; the mailing and nonresponse follow-up procedures; and the weighting, estimation, and data release procedures. From these activities and our deliberations, the panel developed the findings and recommendations presented in this report. 1-B HISTORICAL BACKGROUND 1-B.1 Evolution of the Long-Form Sample The U.S. decennial census, conducted every 10 years beginning in 1790, serves a constitutionally mandated purpose to provide the number of people for each state for reapportionment of the U.S. House of Representatives. A closely linked purpose, to redraw the boundaries of congressional, state, and local legislative districts after each reapportionment, requires that the census provide data at the block level for demographic groups. Beginning as early as 1820, the census obtained additional information beyond the basic head count on the characteristics of the population to respond to the needs of policy makers and the public for a better understanding of the growing new nation. Censuses from 1820 to 1860 included questions on such topics as school attendance, literacy, industry and occupation of employment, and citizenship. Censuses from 1870 through 1930 included a large number of questions asked of everyone. The 1940 census saw the first application of newly developed methods for population sampling to reduce census costs and the burden on the public (Citro, 2000c). In this census, about one-sixth of the total of about 50 questions were asked of only a 5 percent sample. In 1950 about two-fifths of the total of about 50 questions were asked on a sample basis. This approach to obtaining a wide range of information about the population can be termed a “paired strategy,” in which a sample survey that asks a large number of questions is embedded in an enumeration of the entire population on basic characteristics. Prior to 1960, all census data were collected by census enumerators. The 1960 census saw the introduction of the mails to assist the enumeration, with separate short-form and long-form questionnaires. In this and subsequent censuses, the long forms contained the small number of questions asked of everyone—as on the short form—as well as additional questions asked of only a sample (see Citro, 2000b). Long-form-sample sizes varied across censuses. In the 2000 census, the long form was sent to about

OCR for page 13
Using the American Community Survey: Benefits and Challenges one-sixth of housing units overall (17 percent, or about 18 million housing units), although sampling rates varied from 13 to 50 percent depending on the population size of the area. Estimates from the long-form sample were released for areas as small as census tracts and block groups. Estimates for the entire population from the data collected for everyone were released at the individual block level, but this was not the case for long-form-sample estimates, both because the estimates were not sufficiently precise at that level and out of concern to protect individual confidentiality. The long-form questions have changed over time to reflect changing needs for small-area data to implement federal legislation and administer federal programs. In addition to the basic items asked on the short form and depending on how one counts items with multiple parts, the 2000 census long form included 54 sample items about people, covering such topics as marital status, educational attainment, place of birth, citizenship, language spoken at home, English proficiency, ancestry, military service, year moved into residence, various types of disability, responsibility for grandchildren in the home, current and prior year employment status, occupation and industry, transportation to work, and income by type. The 2000 census long form also included 30 sample items about housing, covering such topics as market value of owned home, rent, cost of utilities, characteristics of house or apartment, year structure built, ownership finances, and number of vehicles. These data have been used by the federal government for such purposes as implementation of sections of the Voting Rights Act, allocation of billions of dollars of federal funds to states and localities, assessment of charges of employment discrimination, and planning, monitoring, and evaluating federal programs. They have also been used by state and local governments for fund allocation, program planning and evaluation, facility planning, and economic development and marketing. Private-sector organizations (retail establishments, restaurants, banks, advertising firms, utility companies, health care providers, etc.) have used long-form-sample data for site location, the targeting of advertising and services, workforce development, and the assessment of compliance with government requirements. Finally, researchers have used long-form-sample data to help understand key social processes, such as internal migration and the correlates of poverty (see National Research Council, 2004b:Ch. 2; National Research Council, 1995:Apps. C, D, E, F, G, H, M, for details). 1-B.2 Why Seek an Alternative to the Long-Form Sample? On one hand, the paired strategy used in modern censuses through 2000 of embedding a long-form sample in the basic decennial census had at least three advantages compared with using a separate survey to collect

OCR for page 13
Using the American Community Survey: Benefits and Challenges long-form-type data on a continuous basis: point-in-time reference periods; close agreement for small areas between the sample estimates and the complete counts by age, sex, race, ethnicity, and housing type; and—under some assumptions—cost savings from taking advantage of the census infrastructure. On the other hand, the paired strategy had at least three disadvantages: decreasing relevance of the data the longer the period between Census Day and the time for which estimates are desired; impaired data quality because of the priority given to completing the head count; and infrequent opportunities to revise the questionnaire. In addition, the paired strategy introduced inefficiencies into the census operations and—perhaps—impaired the completeness of the census head count. 1-B.2.a Advantages of the Paired Strategy The long-form-sample data, like all census data, referred to a single point in time, which in 2000 was Census Day, April 1 (even though questionnaires were mailed in mid-March and follow-up operations to complete the enumeration spread out over several months). For some economic characteristics, such as income, the data referred to a single reference period, which was the preceding calendar year. Accompanying this point-in-time reference period was a concept for assigning people to a specific “usual residence,” which was defined as the place where the person lived or stayed most of the time. These concepts were easy for users to understand and work with. In contrast, a continuous survey requires users to become accustomed to somewhat different and more complex concepts of reference periods and residence. In the case of the ACS, the estimates for a calendar year are based on aggregating data over the 12 months of data collection; the reference period is either the time when a household fills out the questionnaire or the preceding 12 months (for income, weeks and hours worked, and some housing costs); and residence is defined using a 2-month residence concept. The paired strategy had the advantage that estimates for head counts and basic demographic characteristics from the long-form sample could be controlled to conform to the full census figures for small areas, using a statistical procedure called a raking ratio adjustment. This procedure reduced sampling and nonsampling error in the long-form-sample data products and produced a high level of consistency between estimates for demographic groups and small areas from the complete enumeration and the sample. Postcensal estimates are used to control the ACS, but they contain more error than the census counts and are not available for very small areas. The “piggy-backing” of the long-form sample on the existing infrastructure for the short-form census may have had the advantage for the Census Bureau of reducing the costs of administering the long form. “Infra-

OCR for page 13
Using the American Community Survey: Benefits and Challenges structure” includes the MAF or comprehensive list of housing units in the United States; the mailout of forms to most households; the follow-up of nonrespondents and the associated local enumeration office structure; and data capture, editing, imputation, and preparation of data products. Under some assumptions, the marginal cost of the long-form sample was primarily that of the additional data capture and processing; hence, the cost savings through use of the paired strategy. Another source of cost savings is that, with the paired strategy, the MAF needed to be updated only once a decade, whereas the ACS requires the MAF to be constantly updated. An ongoing survey such as the ACS requires separate continuous data collection and processing, as well as continuous estimation, tabulation, and publication operations. Moreover, a continuous survey is never likely to obtain as high a mail response rate as the long form (which benefited from the massive publicity surrounding the decennial census) and hence requires more costly field follow-up. 1-B.2.b Disadvantages of the Paired Strategy From the user perspective, a major disadvantage of the paired strategy was that the data collected became less relevant and more out of date the farther one was from Census Day. Typically, because of the need to give priority to the head count processing, the long-form-sample data were not released until 2 years or more after Census Day, and they were not updated for another 10 years. How much error resulted from the use of out-of-date information depended on how quickly the population of an area changed over time. For example, consider estimates of the immigrant (foreign-born) population for the cities of Los Angeles, San Diego, and San Francisco, California, provided by the ACS test surveys conducted in 2000–2004. The data show complex patterns of stability and change within and among the three cities across this short time span—patterns that would be difficult to predict if the only data available were from a long-form sample every 10 years. The percentages of immigrants in the population of all three cities did not change between 2000 and 2004, but the percentages of the immigrant population who arrived after 2000 and who became citizens grew significantly. In Los Angeles, the percentage of people aged 5 and older who spoke a language other than English increased significantly between 2000 and 2004, but in San Diego and San Francisco, this percentage did not change.2 The effects of Hurricanes Katrina and Rita on the populations of New 2 At http://factfinder.census.gov, specify “get data” for the ACS and go to “Data Profiles, Selected Social Characteristics” for a specific year. All web addresses are current as of March 2007.

OCR for page 13
Using the American Community Survey: Benefits and Challenges Orleans and other affected Gulf Coast areas offer a striking example of the way in which long-form-sample data from a census could become irrelevant. Even though increasingly out of date, the 2000 long-form-sample data may have approximated the characteristics of the population in these areas reasonably well through the summer of 2005. However, after the hurricanes hit and so many people fled the area, moved into temporary housing, lost their jobs, or experienced other major changes in their living situations, the 2000 long-form-sample data no longer came close to approximating the numbers and characteristics of the remaining residents. In contrast, in June 2006 the Census Bureau was able to issue a special product from the 2005 ACS for these areas, providing separate estimates for the period January-August 2005 and the period September-December 2005.3 Not only were the long-form-sample products delayed because of the priority given to completing the head count, but also in 2000, even more than in prior censuses, there was not a dedicated effort to collecting long-form information during nonresponse follow-up. Furthermore, in 2000 there was no operation to follow up households that mailed back incomplete long-form questionnaires. Consequently, nonresponse rates were very high in 2000 for many long-form items, particularly those obtained in follow-up operations by temporary, minimally trained field staff. They were high absolutely and in comparison with the 2000 ACS test survey (known as the Census 2000 Supplementary Survey or C2SS), for which the Census Bureau used permanent, highly trained interviewers in nonresponse follow-up (Schneider, 2004:Appendix Tables 1, 2). Measurement error may also have been greater for long-form-type information collected as part of the census than is the case for the ACS. The paired strategy limited the opportunities to revise the questionnaire to respond to emerging data needs or to improve response quality. Although it is unclear how often ACS questions can be revised, the strategy of obtaining long-form-type information in the continuous measurement design of the ACS should allow for additions to the questionnaire to address current issues of interest more frequently than once every 10 years. An important consideration for the Census Bureau is that the infrastructure needed to administer a short-form census can be much more efficient than the infrastructure needed to administer both the short form and the long form. Indeed, the current design for carrying out the short-form-only census in 2010 envisions the use of handheld computing devices 3 Available at http://www.census.gov/acs/www/Products/Profiles/gulf_coast/index.htm. The special ACS estimates were not without problems—the hurricanes disrupted postal service delivery, dislocated the interviewer workforce, and made it difficult to complete interviews from sample households. Moreover, post-hurricane population and housing unit estimates were not available to use as controls for the ACS post-hurricane estimates, so no controls were used.

OCR for page 13
Using the American Community Survey: Benefits and Challenges to collect and transmit data during nonresponse follow-up operations and to assist in managing the large, temporary enumerator workforce. These handheld devices are considered crucial by the Census Bureau to lowering the costs of field data collection by reducing the time needed to find the next address in an enumerator’s daily assignments, by automatically capturing and transmitting data, by reducing the amount of paper to be managed, and by helping to better manage enumerators’ assignments and compensation. Not having to collect long-form data facilitates the use of handheld devices by the census enumerators because many fewer data items must be asked and recorded. Also, not having the additional long-form questions should reduce the amount of paper that the centers responsible for processing the mail returns will need to handle by about 20 percent. These considerations certainly reduce the cost-efficiency argument of having the long-form sample use the short-form infrastructure, although whether they would completely overcome the cost advantages of the paired strategy compared with a separate continuous survey is not clear. Finally, it is possible that dropping the long form from the census will improve the completeness and accuracy of the basic head count. In 2000, adverse publicity about the perceived intrusiveness of questions on the long form helped reduce the mail return rate for long forms, which was 9 percentage points below that for short forms (National Research Council, 2004b:Box 4.1). A short-form-only census, other things being equal, will probably have a higher mail return rate than in 2000. The higher the return rate, the less is the likelihood that people will be missed or doublecounted because they moved between Census Day and the time (several weeks or months later) that census enumerators visit nonresponding households or because the enumerators will find no one at home and obtain information from a neighbor or landlord. 1-B.3 Evolution of the ACS Considerations of data needs and census efficiencies drove the development of the ACS as a replacement for the long-form sample. Several European countries have moved to continuous measurement as well—see Box 1-1. The need for more frequent small-area data on the social and economic conditions of the population was discussed at least as far back as a 1941 proposal by Philip Hauser, then deputy director of the Census Bureau, for an “annual sample census” (see Alexander, 2001). In 1981, Leslie Kish of the Institute for Social Research at the University of Michigan proposed a “continuous measurement” or “rolling sample” design in place of the census, in which one-tenth of the population would be surveyed each year, cumulating the estimates over 1, 2, or more years to increase their precision

OCR for page 13
Using the American Community Survey: Benefits and Challenges BOX 1-1 Continuous Measurement in Three Countries The United States is not alone among developed countries in exploring the advantages of a rolling sample design for some or all of the content of a traditional census. Such investigation in other countries has usually assumed that a rolling design would replace a traditional census and not just long-form-type characteristics and that extensive use would be made of administrative records. France In 2004 France replaced its census with a form of rolling sample design (Desplanques, 2003). Every year a sample is drawn that comprises all households in one-fifth of the smallest communes (local administrative units) together with one-fifth of households in larger-sized communes. The data are cumulated over 5 years; in addition, data from administrative records are used to supplement the yearly samples to produce more reliable 1-year estimates. Germany Germany has not conducted a census since 1987 due to privacy concerns. In its place, Germany conducts a microcensus, involving every year 1 percent of all households in Germany, each of which stays in the sample for 4 years (see http://www.destatis.de/micro/e/micro_c1.htm). The questionnaire includes mandatory and supplementary items. Data from the microcensus and administrative records are used to develop population estimates. Great Britain The national statistical agency of Great Britain has explored the pros and cons of a rolling census. The Office for National Statistics (2003) reached a conclusion that a rolling design could be feasible to implement following the next full census (scheduled for 2011) if population and address registers are sufficiently developed to augment the sample-based data. for small geographic areas (National Research Council, 1995:71).4 Daniel Horvitz (1986) proposed a design that rolled by geographic area: a full census, including short-form and long-form content, would be conducted every year of one-tenth of the nation’s counties. In 1988, Roger Herriot, then chief of the Population Division of the Census Bureau, proposed an ongoing “Decade Census Program” that mixed aspects of the Kish and Horvitz designs. 4 Kish wrote a series of papers on continuous measurement, advocating its use for a variety of purposes, particularly in developing countries (see Alexander, 2001).

OCR for page 13
Using the American Community Survey: Benefits and Challenges After the 1990 census, members of Congress expressed concern about the perceived adverse effect of the long-form questionnaire on the completeness of census response. In 1990, 29 percent of households that received the long form failed to mail back their form, compared with 24 percent of households that received the short form (National Research Council, 2004b:100), and some thought that this differential contributed to the poorer coverage of the population in 1990 in comparison with 1980. Other members of Congress were interested in more frequent estimates for small areas of such long-form-sample statistics as the percentage of school-age children in poverty for use in allocating federal funds to states and localities. Consequently, in 1994, the Census Bureau formed a staff to implement a continuous measurement design similar to that proposed by Kish in a few test sites, so that it could be evaluated as a replacement for the long-form sample in 2000. Renamed the American Community Survey, a questionnaire similar to the long form was fielded in four counties in 1996.5 The decision was made in the mid-1990s to retain the long-form sample in 2000 and to implement the ACS on a gradual basis, so that it could be further tested and compared with the 2000 census results. The goal was to conduct a short-form-only census in 2010 and to fully implement the ACS so that it could provide estimates for small areas that were about as precise as long-form-sample estimates for small areas by accumulating samples over 5 years. However, very early in the development process, rising costs led to a decision to scale back the originally planned sample size of 500,000 housing units per month to a sample of 250,000 housing units per month (National Research Council, 1995:127). The original 4 ACS test sites were gradually increased to include 36 counties in 31 sites for the years 1999–2004 (see U.S. Census Bureau, 2003). In addition, the C2SS was fielded as an experiment in more than one-third of U.S. counties by using the monthly ACS data collection design and a questionnaire very similar to the long form. Cumulated over the 12 months of 2000, the C2SS, including the test sites, obtained responses from about 587,000 housing units; it demonstrated the feasibility of conducting an ACS nationwide and at the same time as the decennial census (U.S. Census Bureau, 2001). Similar nationwide test surveys were fielded in 2001–2004. In January 2005, full implementation of the ACS was inaugurated for the first month’s sample of 250,000 housing units; and in January 2006, group quarters residents were added to the ACS. Plans are to proceed with a short-form-only census in 2010. 5 The late Charles (“Chip”) Alexander played a pivotal role in designing the ACS and moving it toward full implementation (see, for example, Alexander, 1993, 1997, 1998).

OCR for page 13
Using the American Community Survey: Benefits and Challenges 1-C ISSUES FOR THE PANEL Two basic features of the ACS data products raised key questions for the panel’s consideration: (1) the plan to use multiyear estimates from aggregating monthly samples from the ACS to replace point-in-time estimates from the decennial census long-form sample and (2) the plan to use census-based population estimates at the level of counties (or groups of small counties) for July 1 of each year to calibrate the sample-based ACS estimates. The purpose of using the population estimates, which are developed from the previous census updated with births, deaths, and migrant flows from administrative records, as controls is to reduce the effects of sampling error and compensate for any incompleteness of coverage in the population surveyed in the ACS. There are also separate controls for housing units. The use of multiyear averages is central to the ACS design. The question it raises is the extent to which users can easily apply the ACS data products to the important and varied uses that, until now, were met through the long-form-sample data products for the census year. What does it mean to have a 5-year estimate for an area of, say, the poverty level or the average length of time to commute to work cumulated from 60 months’ worth of data? How should users interpret differences in estimates for the same geographic area, such as a large county, that will be available from data cumulated over 1, 3, and 5 years? Turning to the detailed procedures for producing multiyear estimates, does the Census Bureau’s plan to weight the data to the average population over the period represent the optimal approach, or are there other approaches that might on balance have better properties? The decision to use controls for counties (or groups of small counties) for specific population groups (defined by age, sex, race, and ethnicity) for July 1 each year raises several questions. One question concerns the effect of controlling the ACS monthly samples spread over a year to point-in-time population estimates that are updated from April 1 population counts from the last census. Another question concerns whether the county population estimates by age, sex, race, and ethnicity are of adequate quality to be used for this purpose. Might, for instance, controlling to population estimates at higher levels of geographic aggregation or using fewer population groups as controls offer advantages over the current plan? In addition to these two primary areas of investigation, the panel needed to examine the ACS as a data collection and production system to answer questions concerning the functionality and usability of ACS data products. This examination led the panel to address four areas of ACS operations: (1) sampling for housing units, including initial sampling from the MAF and subsampling of nonrespondents for follow-up by computer-assisted personal interviewing; (2) data collection for housing units,

OCR for page 13
Using the American Community Survey: Benefits and Challenges including the mode of collection (mail, telephone, personal visit) and residence rules; (3) sampling and data collection for group quarters; and (4) treatment of the data, including disclosure avoidance, collapsing of tables to improve precision, inflation adjustments for income and housing costs, the population universe and geographic areas for tabulations, and data quality review. A last broad issue concerns the ultimate role of the ACS in the federal statistical system. Since the ACS has been developed to replace—and improve on—the decennial long-form sample, it is first necessary to assess the quality and usability of the ACS data products in comparison with the long-form-sample data products. The panel thinks, however, that focusing on these comparisons limits one’s perspective concerning the ultimate utility of the ACS. Instead of viewing the ACS data products as simply analogous to or as a substitute for decennial census long-form-sample data products, the panel thinks that the ACS should be viewed as an entirely new and unique source of information to support public and private decision making. ACS data can be used as input for analyses that would not be feasible with long-form-sample data products. In that context, the panel evaluated the various characteristics of the ACS and its data products with a longer term view of what the potential for the ACS will be over time, and how it can be helped toward achieving its full promise as a key component of the federal collection of individual and household demographic, social, and economic data. 1-D OVERVIEW OF THE REPORT Following this introduction, the panel’s report has three parts: using the ACS (Part I), technical issues (Part II), and priorities for user outreach and continued development of the ACS (Part III). Chapters 2 and 3 in Part I are addressed to users of long-form-sample data who want to use the ACS for their applications. Chapter 2 reviews key features of the ACS design; the major advantages of the survey in terms of timeliness, frequency, and quality of data; and the major challenges of using the ACS data in terms of period estimates replacing point-in-time estimates and high levels of sampling error for small areas. This chapter is essential reading for users who will work with the ACS data. It ends with an assessment of the usefulness of the ACS. Chapter 3 provides guidance for applying the ACS data. Through examples of key applications that currently use long-form-sample data, the chapter considers how users can work with the various ACS products for those applications and the considerations they need to take into account in deciding which products to use for their applications. Both Chapters 2 and 3 take as given the Census Bureau’s announced plans for ACS data prod-

OCR for page 13
Using the American Community Survey: Benefits and Challenges ucts. Both chapters are written to be as user-friendly as possible, although, given the complexity of the ACS, they will be most helpful to two groups: people who expect to use the ACS for repeated, in-depth applications and people who expect to serve as intermediaries helping other users with their applications. The next three chapters in Part II are addressed to technical users and the Census Bureau. They critically review key features of the ACS design, operations, and data products and offer recommendations for areas of further research and possible modification in the future. Chapter 4 addresses features of the ACS sample design and operations that are particularly relevant to the quality of the data and hence their usability for various applications. Chapter 5 reviews the procedures to weight the 1-year data so that totals agree with the Census Bureau’s population estimates for major demographic groups, as well the Bureau’s housing estimates, in large counties and groups of smaller counties. The chapter also discusses adjustments that are made to the data to account for nonresponse. Chapter 6 reviews the construction, interpretation, and possible alternative methods for producing multiyear period estimates. Chapter 7 in Part III brings the user and technical strands together: it reviews and makes recommendations for three key areas of continuous research and development for the ACS. The three are (1) education and outreach activities to various user communities to help ease the transition from the census long-form-sample data products to those from the ACS and to ensure a continuous feedback loop between the Census Bureau and data users; (2) priorities for continuous methodological and operational improvement of the ACS; and (3) a vision of the future in which the ACS contributes in new and innovative ways to expanding information on the nation’s people and communities. The appendixes include a list of abbreviations and acronyms used in the report (Appendix A) and two papers written for the panel by F. Jay Breidt: Appendix B is on population controls and Appendix C is on multiyear period estimates. The report concludes with biographical sketches of panel members and staff in Appendix D.

OCR for page 13
Using the American Community Survey: Benefits and Challenges This page intentionally left blank.