Appendix C

Additional Frameworks Relevant for Federal Statistics

With the expansion of data forms and supply, and with increasing regulation in data use and access, there has been concurrent and consecutive guidance on the professional conduct of statistics and data science, both nationally and internationally. The most significant of these are included in this appendix as a way to recognize shared professional values and norms in the production and dissemination of statistics.

Each of these guidances was developed for somewhat different purposes and audiences. It is therefore reasonable to anticipate some differences in alignment. Nonetheless, despite these differences of audience and purpose, others have noted broad alignment of guidelines overall (Park & Tractenberg, 2023; Tractenberg & Park, 2023). Even taking into account the time spanned across releases and geographies, these guidelines signal a strong, common understanding of ethical statistical practice and its importance. At an elemental level, areas of alignment were strongest pertaining to credibility, transparency, and trust. Elements of independence, relevance, accountability, and improvement were also common, but to a lesser degree. (Park & Tractenberg, 2023; Tractenberg & Park, 2023).

Several of the descriptions that follow were drawn from the overview of national and international guidelines for the practice of official statistics provided in (Park & Tractenberg, 2023; Tractenberg & Park, 2023), as noted.

U.S. GUIDANCE FOR ETHICAL DATA AND STATISTICAL PRACTICE

Interagency Council on Statistical Policy Principles for Modernizing Production of Federal Statistics

A shift from reliance on surveys for primary data collection to reliance on surveys as a complement to already existing data, either in-house or otherwise available to an agency, is essential to the modernization of the federal statistical system (National Academies of Sciences, Engineering, and Medicine [NASEM], 2017c,e).

The use for statistical purposes of nonstatistical or integrated data—also known as blended, hybrid, or combined data—requires investment in data acquisition and development of standards. First, statistical agencies and components must be able to access alternative data sources. A second major challenge is measuring and transparently communicating the quality of statistical information derived from nonstatistical or integrated statistical and nonstatistical information (NASEM, 2017c,e).

In 2018, the Interagency Council on Statistical Policy (ICSP) issued the principles for work on integrated data listed in Box C-1.¹ These principles are intended to guide the Federal Committee on Statistical Methodology (FCSM) and to establish priorities for research and ongoing work by statistical agencies to advance the use of nontraditional information and data to produce statistical information. The work by the ICSP and the FCSM fits more broadly under Office of Management and Budget’s development of a Federal Data Strategy.

In 2023–2024, the Committee on National Statistics (CNSTAT) prepared a series of consensus reports providing a vision for a new data infrastructure—one that relies heavily on the appropriate use and management of blended data to inform longstanding policy issues with greater efficiency and lower public burden (NASEM, 2023b,c, 2024c). These themes continued in several other CNSTAT reports (NASEM, 2016b, 2024b).

Federal Data Ethics Tenets

The U.S. Federal Data Strategy was issued in 2019 to advance access, interoperability, and utility of federal data. Its 2020 Action Plan called for the development of a Data Ethics Framework to help agency employees, managers, and leaders make ethical decisions as they acquire, manage, and use data throughout the data lifecycle. The intended audience for this framework is all federal agencies involved in the data lifecycle—extending

American Statistical Association Ethical Guidelines

The American Statistical Association (ASA) is one of the oldest continuously operating professional organizations for statistical practice in the world. It was founded in the United States in 1839 and has had international membership ever since. It was established to promote the practice and profession of statistics both in the United States and worldwide. The ASA Ethical Guidelines for Statistical Practice represent the ethical practice standard for statistics and data science (Tractenberg, 2022a,b). These ethical practice standards were originally adopted by the ASA in 1995 (Hogan & Steffey, 2014) and have been periodically updated since 2016. The most recent update (American Statistical Association, 2022) includes 60 items organized under eight principles (see Box C-3) and a 12-item appendix specific to organizations and institutions. The next anticipated revision will occur in 2027.

The ASA Ethical Guidelines are comprehensive and actionable—not aspirational. Several studies have shown alignment of the ASA Ethical Guidelines with ethical guidance from the Royal Statistical Society (Royal Statistical Society, 2014; Tractenberg, 2020) and the Association for Computing Machinery (Gotterbarn et al., 2018; Tractenberg, 2022a), as well as U.S. law (“Foundations for Evidence-Based Policymaking Act of 2018,” 2019) and policy (General Services Administration, 2020; Office of Management and Budget, 2014b). (Also see Park & Tractenberg, 2023.) Given this strong alignment, some experts recommend the ASA Ethical Guidelines as a useful, actionable tool for communicating and monitoring ethical statistical practice to practitioners using statistics within a range of disciplines, and for audiences in the United States. Tractenberg (2022a, 2022b) refers

American Association for Public Opinion Research Code of Professional Ethics and Practices

The American Association of Public Opinion Research (AAPOR) reissued its Code of Professional Ethics and Practices (Code) in 2021. The Code is intended to support sound and ethical practice in the conduct of public opinion and survey research and promote the informed and appropriate use of research results. The Code is applicable to all public opinion and survey researchers, regardless of AAPOR membership. (See Box C-5 for the Code’s list of contents.)

To summarize, the AAPOR Code requires “[…] the highest standards of scientific competence, integrity, accountability, and transparency in designing, conducting, analyzing, and reporting our work, and in our interactions with participants (sometimes referred to as respondents or subjects), clients, and the users of our research.”³ The Code is notable with its explicit reference to human rights, as well as a rejection of work inconsistent with its tenets: “[w]e pledge to act in accordance with principles of basic human rights in research. We further pledge to reject all tasks or assignments that would require activities inconsistent with the principles of this Code.”⁴

Office of Science and Technology Policy Blueprint for an AI Bill of Rights

In October 2022, the Office of Science and Technology Policy (OSTP) issued the Blueprint for an AI Bill of Rights. The Bill provides “[…] five principles that should guide the design, use, and deployment of automated systems to protect the American public in the age of artificial intelligence.”⁵ See Box C-6. The Bill is accompanied by a handbook From Principles to Practice, intended to assist persons seeking to include these protections into practice, including detailed steps for implementing these principles in the technological design process.

INTERNATIONAL GUIDELINES FOR OFFICIAL STATISTICAL PRACTICE

United Nations Fundamental Principles of Official Statistics

The United Nations Statistics Division website notes that the need for a set of principles governing official statistics became apparent at the end of the 1980s, when countries in Central Europe began to change from centrally planned economies to market-oriented democracies. It was essential

In 2015, the United Nations Statistical Commission issued implementation guidelines for the Fundamental Principles.⁷ These guidelines list several actions or activities that a statistical agency is advised to take into account when aiming to improve the practical and effective implementation of a certain principle or when developing a certain principle further. Concrete as well as practice-orientated examples (good practices) complement these recommended actions. Part II of the implementation guidelines contains recommendations on how to ensure a high level of independence for national statistical systems. These guidelines differentiate between various forms of independence (such as institutional, professional, and scientific independence) and recommend good practices in order to ensure independence.

International Statistical Institute Declaration on Professional Ethics

The International Statistical Institute (ISI) is a nonprofit, nongovernmental organization established in 1885 with individual and institutional members in over 150 countries. Its main objective is to promote the understanding, development, and good practice of statistics worldwide to exchange advances in statistical knowledge and best practices and by creating opportunities to network.

The ISI Declaration on Professional Ethics was first issued in 1985, and subsequently revised in 2010 and, most recently, in 2023 (see Box C-8). The aim of the Declaration is “to enable the statistician’s individual ethical judgments and decisions to be informed by shared values and experience. [...]” It “seeks to document widely held principles of the statistical profession and to identify the factors that obstruct their implementation” and offers an aspirational framework to guide this work.

Rather than implementation guidelines per se, the ISI provides a description of roles and strategies with regard to professional ethics. These can be summarized as ways to communicate support to professional statisticians, note concern to national authorities and international agencies, and collaborate on joint communications with other professional societies where possible to amplify messages.

European Statistics Code of Practice

The European Statistical System Committee (ESSC) is a partnership in which the European Union Statistical Authority (Eurostat) and the national statistical authorities of each European Union member state and each European Free Trade Association state cooperate. The joint mission of these

In Article 3 of the OECD Convention, members agreed to “furnish the Organisation with the information necessary for the accomplishment of its tasks.” The quality of statistics is fundamental for the quality of evidence-based analytical work of the OECD and for the quality of statistical publications and databases that it produces. Adopted by the OECD Council in 2015 on the proposal of the OECD Committee on Statistics and Statistical Policy (CSSP), the OECD Good Statistical Practice Recommendations provide a detailed blueprint for a sound and credible national statistical system. Each of the 12 specific recommendations (see Box C-10) is supplemented by a set of indicative good practices, which together provide a framework for examining the implementation of the Good Statistical Practice Recommendations. The Recommendations themselves are presented as a series of yes/no checkboxes oriented to the concerns of national statistical offices, rather than offering guidance for ethical statistical practice at the individual practitioner level.

The OECD Good Statistical Practice Recommendations also provide a benchmark against which the national statistical system of countries can be assessed. They constitute a tool for self-assessment of non-members, which can facilitate the identification of needed improvements in their national statistical systems. Individual assessments of national statistical systems could take several forms: a simple self-assessment, an evidence-based self-assessment, or a peer review by the CSSP.

Periodic reports to the Committee on the implementation of the Recommendations are based on information gathered through a number of tools, including a survey among Adherents to the Recommendations, a peer review of national statistical systems, and self-assessment questionnaires provided by Adherents. Reviews describe good practices and common challenges. The next report is anticipated in 2025.

GENERIC STATISTICAL BUSINESS PROCESS MODEL

The Generic Statistical Business Process Model (GSBPM) of the United Nations Economic Commission for Europe was first developed in 2008 and most recently updated in 2019 (version 5.1). The model is designed to enable statistical agencies to describe production processes in a coherent way, compare processes within and among organizations, and make better decisions on production systems and allocation of resources. The GSBPM (shown in Figure C-1) describes and defines the set of business processes needed to produce official statistics. It provides a standard framework and harmonized terminology to help statistical organizations modernize their statistical production processes, as well as to share methods and components. The GSBPM can also be used for integrating data and metadata standards, as a template for process documentation, for harmonizing statistical computing infrastructures, and to provide a framework for process quality assessment and improvement.