FEDERAL STATISTICS HAVE BEEN INTEGRAL to the growth and development of the United States since the adoption of the U.S. Constitution: Article 1, Section 2, mandates a decennial census to provide population counts for periodic reallocation of seats in the House of Representatives. Today, federal statistics undergird the allocation of billions of dollars of federal funding to states, localities, and other entities, drive financial markets and decision making, contribute to policy debates in many areas of public interest, and support valuable scientific research.
Historically, in the United States, as particular policy concerns have become salient at the federal level, Congress has established statistical agencies to provide the information needed to address them. As a consequence, the federal government has a large number of statistical agencies, as well as agencies with statistical programs, and a statistical coordination function in the U.S. Office of Management and Budget (OMB). The principles and practices for a federal statistical agency that are the subject of this report pertain to individual statistical agencies as separate entities in a decentralized system for providing federal statistics.1
This report is the sixth edition of Principles and Practices for a Federal Statistical Agency, first published in 1992 by the Committee on National Statistics (CNSTAT) at the National Academies of Sciences, Engineering, and Medicine. The publication draws on CNSTAT’s many studies of specific agencies, programs, and topics. It has proved helpful to Congress, OMB, federal statistical agencies and others about what constitutes an effective
1 This report neither comments on the advantages or disadvantages of the decentralized U.S. system nor compares it with other models for organizing government statistics (see Norwood, 1995, for a comparison). “Statistical agencies” refers to the nation’s 13 principal statistical agencies: see “Definition of a Federal Statistical Agency” below.
and credible statistics entity.2 Beginning with the second edition in 2001, CNSTAT has updated the document every 4 years to provide a current edition to newly appointed cabinet secretaries and other personnel at the beginning of each presidential administration or second term.
The Committee on National Statistics had the following statement of work for this sixth edition:
In response to recurring requests for advice on what constitutes an effective federal statistical agency, CNSTAT issued the first edition of Principles and Practices for a Federal Statistical Agency (P&P) in 1992. In early 2001, 2005, 2009, and 2013, CNSTAT issued the second, third, fourth, and fifth editions, respectively, which reiterated the basic principles for federal statistical agencies, revised and expanded the discussion of some of the practices for an effective statistical agency, and updated the discussion with references to recent reports by CNSTAT and others. Changes in laws, regulations, and other aspects of the environment of federal statistical agencies over the past 4 years [see Appendix A] warrant preparation of a sixth edition, which a CNSTAT committee will prepare for release in early 2017.
The committee distinguishes between “principles,” which are fundamental and intrinsic to the concept of a federal statistical agency, and “practices,” which are ways and means of making the basic principles operational and facilitating an agency’s adherence to them. Other documents that present principles, practices, or both for statistical agencies include the Fundamental Principles of Official Statistics (United Nations Statistical Commission, 2014), first promulgated by the United Nations Statistical Commission in 1994 and endorsed by the United Nations General Assembly with a revised preamble in 2014; the European Statistics Code of Practice for the National and Community Statistical Authorities (European Statistical System Committee, 2011), first issued in 2005 and revised in 2011; and Statistical Policy Directive No. 1, Fundamental Responsibilities of Federal Statistical Agencies and Recognized Statistical Units (U.S. Office of Management and Budget, 2014).3
This sixth edition presents and comments on four basic principles that statistical agencies must embody to carry out their mission fully:
- They must produce data that are relevant to policy issues.
- They must achieve and maintain credibility among data users.
- They must achieve and maintain trust among data providers.
- They must achieve and maintain independence from the appearance and reality of political or other undue external inﬂuence in developing, producing, and disseminating statistics.
This edition also discusses 13 important practices that are the means for statistical agencies to implement the four principles. These practices include authority for an agency to protect its independence, use of multiple data sources for statistics that meet user needs, openness about sources and limitations of the data provided, respect for privacy and autonomy of data providers, protection of the confidentiality of providers’ information, a commitment to quality and professional standards of practice, a strong internal and external evaluation program, and coordination and collaboration with other statistical agencies.
Having high-quality information from statistical agencies that can address important issues as they arise is essential to inform public and private decisions, to design, monitor, and evaluate effective public policies,4 and to support research that contributes to scientific knowledge and evidence-based policy making. To best serve these needs, this sixth edition emphasizes the importance of statistical agencies’ actively seeking out new ways of using multiple data sources (such as administrative records, private-sector datasets, and other sources in addition to surveys) as the basis for their statistical products. The goal is to enable agencies to maintain and improve the credibility, relevance, accuracy, timeliness, and cost-effectiveness of their data in the face of the significant challenges they now face. Chief among these challenges are the increasing perceptions of burden and unwillingness to respond to traditional survey inquiries by people and organizations, the correspondingly higher costs of obtaining adequate responses to survey inquiries, and constrained budgets for statistical activities (see Citro, 2014a; Groves, 2011; National Academies of Sciences, Engineering, and Medicine, 2017a; Prewitt, 2010:7–16; Singer, 2016). In facing these challenges, the growing availability of multiple data sources and advances in computing technology and methods for combining data offer many opportunities for more complete and timely statistics.
Unlike previous editions, this sixth edition presents the principles and practices in separate parts (II and III), with commentary on them included. The rest of this Part I covers four topics: the definition of a statistical agency; reasons for establishing a statistical agency or adding responsibilities to an existing agency; a brief history of the U.S. statistical system; and a discussion of the value of federal statistics. Throughout the online version of the report, hyperlinks to other sections and to references are provided wherever possible to assist the reader. The two appendixes update material included in previous editions and add new information to help orient readers; their contents are listed in this document, and they
are available in full in the on-line edition.5Appendix A summarizes the history and current status of key legislation and regulations that affect federal statistical agencies, such as the Confidential Information Protection and Statistical Efficiency Act of 2002; OMB statistical policy directives; and Office of Science and Technology Policy memoranda on scientific integrity and public access to federally sponsored research and data. Appendix B reviews the organization of the U.S. federal statistical system, providing information about each recognized statistical agency and statistical unit and the coordinating function in OMB.
Although focused on federal statistical agencies, many of the principles and practices articulated here likely also apply to statistical activities elsewhere, such as in federal policy, evaluation, research, and program agencies, in state and local government agencies, and in other countries. The principles and practices in this report remain guidelines, not prescriptions. CNSTAT intends them to assist statistical agencies and to inform legislative and executive branch decision makers, data users, and others about the characteristics of statistical agencies that enable them to serve the public good.
5 Available at doi: https://doi.org/10.17226/24810. The National Academies Press website (www.nap.edu) provides free access to all published reports of the National Academies of Sciences, Engineering, and Medicine in electronic formats.