EXECUTIVE SUMMARY
Two major criticisms were levelled against the 1990 census: (1) unit costs increased significantly, continuing a trend that began with the 1970 census; (2) the problem of differential undercount by race persisted and even worsened, despite a large investment in coverage improvement programs (see, e.g., U.S. General Accounting Office, 1992). In response to these criticisms, the Census Bureau is considering an unprecedented level of innovation for the 2000 census.
In this interim report we concentrate on those aspects of census methodology that have the greatest effect on these two primary objectives of census redesign: reducing differential undercount and controlling costs. Therefore, we focus on processes for the collection of data, the quality of coverage and response that these processes engender, and the use of sampling (and subsequent estimation) in the collection process.
Census data collection involves four key steps: (1) the construction of an address frame; (2) an initial process to obtain responses that can be linked to the address frame; (3) a follow-up process to obtain responses from those not covered in the initial process; and (4) a coverage assessment process that estimates the size of the population not covered through the initial and follow-up processes. In the 1990 and earlier censuses, the first three steps led to the official census estimates; whether or not to incorporate the estimates from the fourth step into the official census estimates became the ''adjustment issue.'' For the 2000 census, the Census Bureau is proposing a fundamentally different approach, called a "one-number census." The one-number census describes an approach that regards this fourth step as an integral part of the census process that leads to the official estimates.
The design of a census data collection process in essence amounts to deciding which methods of identification, enumeration, response, and coverage improvement should be applied at each of the steps; whether sampling methods (and the corresponding estimation methods) should be used at any of the four steps; and if sampling methods are used, which methods and at which steps. These decisions have to be based on information about the effectiveness and costs of the various alternative methods. The 1995 census test should be a prime source of such information.
In this report we present 35 recommendations that address a broad range of issues with varying degrees of complexity and urgency. Below we present the recommendations organized by the primary intended audience and the associated time frame. (The numbering below follows that of the body of the report.) All but 2 of our recommendations are directed to the Census Bureau, 15 are specifically directed toward the 1995 census test, and 18 are directed toward longer range issues.
ADDRESS LIST DEVELOPMENT AND USE OF ADMINISTRATIVE DATA
Of the two recommendations not directed to the Census Bureau, one calls for congressional action to facilitate cooperative efforts at address list development.
Recommendation 1.3: Congress should enact legislation that permits the sharing of address lists between the Census Bureau and the U.S. Postal Service for the purpose of improving the Census Bureau's master address file.
The second recommendation asks the Office of Management and Budget to assume greater responsibility for statistical uses of administrative records.
Recommendation 4.1: The Statistical Policy Office in the Office of Management and Budget should recognize statistical uses of administrative records as one of its major areas of responsibility and should assume an active role in facilitating more effective working relationships between statistical and program agencies and in tracking relevant legislation.
ONE-NUMBER CENSUS
One key message is that the dual objectives of reducing the differential undercount and controlling costs will require expanded use of sampling and statistical estimation. This theme is prominent in the recommendations regarding the development and testing of coverage measurement methods.
Recommendation 2.3: We endorse the Census Bureau's stated goal of achieving a one-number census in 2000 that incorporates the results from coverage measurement programs, including programs involving sampling and statistical estimation, into the official census population totals. We recommend that research on alternative methodologies continue in pursuit of this goal.
THE 1995 CENSUS TEST
Many of our recommendations address the 1995 census test, including preliminary research that would inform the design of the test.
Coverage Measurement
Recommendation 2.4: Before final design of the 1995 census test, the Census Bureau should critically evaluate the SuperCensus method of coverage
measurement by using 1990 data to learn whether adequately precise coverage estimates are possible using ratios to the housing base.
Recommendation 2.5: Development and testing methodology for the Post-Enumeration Survey (PES) should continue in parallel with other methods until another method proves superior in operational tests. All methods still under consideration—including the PES—should be evaluated critically against common criteria.
Sampling
We call for experimentation in the 1995 census test with the use of sampling to follow up people who do not respond to the initial mail questionnaire and to collect additional information on the census form.
Recommendation 2.1: The Census Bureau should continue research on nonresponse follow-up sampling and truncation, including consideration of a combined strategy with a truncated first stage and sampling during a second stage of follow-up. Evaluation should consider effects of the nonresponse follow-up design on costs and on variance at a variety of geographic levels, from states to small areas.
Recommendation 2.7: The Census Bureau should continue research on possible matrix sampling designs, using the 1990 census data to simulate tabulations and crosstabulations. Design(s) that appear most promising should be tested in 1995 to permit evaluation of their performance in combination with other census design features under test.
Questionnaire Design and Telephone Follow-Up
We make several recommendations with respect to questionnaire design and implementation in the 1995 census test, including use of the telephone to contact persons who do not respond to the initial mailing.
Recommendation 3.1: At this time, the Census Bureau should not initiate any further large-scale experiments designed to improve the initial mail response rate. Instead, response improvement research should now consolidate findings from research conducted to date in order to design experiments for the 1995 census test. The primary objective of these experiments should be to identify optimal field procedures that combine features such as advance notification, replacement questionnaires, and telephone follow-up.
Recommendation 3.2: The prospect of having telephone numbers for a large percentage of households in the 2000 census is a potentially important development that should be explored in the Census Bureau's 1995 test—for example, by using the telephone for reminder calls and nonresponse follow-up.
Recommendation 3.12: When developing and applying residence rules, the Census Bureau should consider both the need to accurately enumerate diverse household structures and the potential for mode effects when an instrument is implemented in both self-administered and interviewer-administered forms. In particular, the Census Bureau should simultaneously develop enumeration forms designed for self-administration and telephone administration for use in the 1995 census test. The comparability of these forms should subsequently be evaluated on the basis of 1995 census test results.
Outreach and Promotion
The panel believes that greater attention should be given to census outreach and promotion and to enumeration methods targeted at historically hard-to-count segments of population. In particular, testing of candidate programs and methods should take place in 1995.
Recommendation 3.5: The Census Bureau should establish an ongoing research and development program for decennial census outreach and promotion. The 1995 census test provides an excellent opportunity to conduct and evaluate promising media campaigns and local outreach programs.
Recommendation 3.8: The Census Bureau should consider developing an extensive network of relations between field offices and local community resources. This infrastructure would be maintained in continuous operation between and during census years. The Census Bureau should develop and implement pilot programs in conjunction with the 1995 census test in order to gather information about the potential costs and benefits of a large-scale local outreach program.
Recommendation 3.10: In the 1995 census test, the Census Bureau should evaluate specific measures and procedures that might improve the enumeration of historically undercounted populations. Candidates for study in 1995 should include a larger repertoire of foreign-language materials (both written and audio), more aggressive hiring of community-based enumerators, and greater flexibility in the timing of enumeration (i.e., contact during evenings and weekends). In particular, the Census Bureau should examine the efficacy of moving census day to the middle of the month.
Administrative Records
We note the importance of testing methods that could expand the use of administrative records for statistical purposes. This research will require cooperation between the Census Bureau and agencies that maintain relevant administrative record systems.
Recommendation 2.2: The Census Bureau should study in the 1995 census test the use of administrative records during nonresponse follow-up as a way to reduce the need for conventional follow-up approaches.
Recommendation 4.3: As part of the 1995 census test, the Census Bureau should construct an administrative records database for the test sites.
Recommendation 4.4: The Census Bureau should establish the testing of record linkage procedures as an important goal of the 1995 census test.
Recommendation 4.5: In preparation for uses of administrative records in the 1995 census test, detailed negotiations between the Census Bureau and the other relevant agencies should begin immediately, with the involvement of the Statistical Policy Office of the Office of Management and Budget (see also Recommendation 4.1).
Address List Development
The construction of an address list is a central element in decennial census operations, and the panel believes the potential benefits are sufficient to justify development, starting in fiscal 1994, and maintenance of a continuously updated address file linked to a geographic database.
Recommendation 1.1: The Census Bureau should continue aggressive development of the TIGER (topologically integrated geographic encoding and referencing) system, the master address file (MAF), and integration of these two systems. TIGER/MAF updating activities should begin in fiscal 1994 and should concentrate first on the sites selected for the 1995 census test.
THE 2000 CENSUS
In addition to recommendations for the 1995 census test, two recommendations refer explicitly to the 2000 census—one proposing a goal for coverage measurement, the other urging consideration of available telephone technology. A third recommendation concerns organizational change to facilitate management of decennial
census outreach and promotion.
Recommendation 2.6: Whatever coverage measurement method is used in 2000, the Census Bureau should ensure that a sufficiently large sample is taken so that the single set of counts provides the accuracy needed by data users at pertinent levels of geography.
Recommendation 3.7: The Census Bureau should investigate developing a menu-driven touchtone call routing system for the 2000 census that gives callers to the Census Bureau's toll-free help line quicker access to the specific assistance they want.
Recommendation 3.3: The Census Bureau should assign overall responsibility for decennial census outreach and promotion to a centralized, permanent, and nonpartisan office. The Census Bureau should consider expanding the mission of the extant Public Information Office to include this charge. Evaluation of outreach and promotion programs should be conducted by an independent unit within the Census Bureau.
FURTHER RESEARCH
The remainder of the recommendations endorse topics for further research throughout this decade. This research might inform the 1995 census test, the 2000 census, or census design beyond 2000. These recommendations cover methods for linking records from one or more sources—that is, mail questionnaires, telephone or personal interviews, or administrative data systems—aimed at improving census accuracy by reducing both omissions and erroneous enumerations, the use of sampling and statistical estimation, outreach and promotion, racial and ethnic classification, the long-term use of administrative records, and continuous data collection.
Record Linkage
Recommendation 1.2: The Census Bureau should aggressively pursue its research program on record linkage.
Sampling and Statistical Estimation
Recommendation 2.8: The Census Bureau should vigorously pursue research on statistical estimation now and throughout the decade. Topics should include nonresponse follow-up sampling, coverage estimation, incorporation of varied information sources (including administrative records), indirect estimation for
small areas, and matrix sampling.
Recommendation 2.9: The Census Bureau should develop methods for measuring and modeling all sources of error in the census and for showing uncertainty in published tabulations or otherwise enabling users to estimate uncertainty.
Outreach and Promotion
Recommendation 3.4: The Census Bureau should commit the resources necessary to develop and implement customized, local outreach programs to target the traditionally undercounted ethnic minorities. The Census Awareness and Products Program (CAPP) should be expanded and sustained on an ongoing basis, so that it can serve as the primary vehicle for the design and implementation of these outreach programs.
Recommendation 3.6: The Census Bureau should evaluate the use of the Advertising Council to conduct the census media campaign. The Census Bureau should consider the alternatives of working directly with local and regional agencies, undertaking paid media research, and supplementing pro bono advertising with paid advertising in hard-to-enumerate localities.
The Differential Undercount and Racial and Ethnic Classification
Recommendation 3.9: The Census Bureau should conduct further comparative studies of hard-to-enumerate areas, focusing on those parts of the country where three phenomena coincide: a shortage of affordable housing, a high proportion of undocumented immigrants, and the presence of low-income neighborhoods.
Recommendation 3.11: The Census Bureau should consider a major program of research in cognitive anthropology, sociology, and psychology that will comprehensively examine the issue of racial and ethnic identity. This research would contribute to the development of more acceptable racial and ethnic identification questions. In particular, the Census Bureau should consider experimenting with allowing people to select more than one race category in the 1995 census test.
Use of Administrative Records
Recommendation 4.2: The Census Bureau should initiate a systematic process of consultation and research to explore the attitudes of the public, political representatives, and other opinion leaders about the use of administrative records as an integral part of the census. Previous consultations and existing research, such as the yet-to-be-released 1990 Taxpayer Opinion Survey, should be taken into account.
Recommendation 4.6: The Census Bureau should establish a formal program of long-range research and development activities relating to expanded use of administrative records for demographic data.
Continuous Data Collection
Recommendation 5.1: The Census Bureau should continue to explore the feasibility of a continuous measurement component to the 2000 census.
Recommendation 5.2: The Census Bureau should establish a formal set of goals for a continuous measurement program. The Census Bureau should then establish a research plan to determine the extent to which these goals are achievable.
Recommendation 5.3: The Census Bureau should undertake an extensive and systematic evaluation of the benefits from having more frequent census data available for both large and small geographic areas.
Recommendation 5.4: The goals for a continuous measurement program (see Recommendation 5.2) should include phasing in the continuous measurement program during the latter half of the decade prior to the relevant census year.
Recommendation 5.5: As part of its research into the feasibility of and methods for implementing a continuous measurement program, the Census Bureau should undertake a thorough study of the consequences of changes in the instrument over time, as well as changes in mode effects. A plan must be established for incorporating the effects of such changes into the cumulated estimates and into the time series produced by the continuous measurement program.
