Daniel Elchert, ASA Science Policy Fellow

Statistics can be described as the science of learning from data and of measuring, controlling, and communicating uncertainty. In practice, a statistician collects and analyzes information using a particular approach to make good inferences in response to some question. Sound data collection and analysis, in turn, help drive decisions supported by evidence.

Foundational to the practice of statistics, and as I believe to democracy writ large, is an objective, trustworthy, and independent federal data infrastructure, or a publicly financed system of statistical agencies that inform the public on topics such as labor, health, education, agriculture, housing, finance, and transportation. But what is our federal data infrastructure, exactly? And when and how might statisticians go about advocating for this system as a means to leverage data for the public good?

The Data Infrastructure

In the United States, statisticians were among the first scientists to go about establishing a profession. It was 1839 when the newly formed American Statistical Association was given the task of “collecting, preserving, and diffusing statistical information,” according to R. L. Mason, J. D. McKenzie, and S. J. Ruberg in The American Statistician article, “A Brief History of the American Statistical Association, 1839–1989.” Using their unique statistical skills, ASA members soon began proposing reforms to the upcoming 1850 census.

These and other early interactions with the federal government signified a burgeoning social contract between American statisticians and the nascent republic’s citizenry; statisticians were seen as experts of their domain and, in turn, entrusted to collect, analyze, and report official data in agriculture, commerce, education, and health.

A federal data infrastructure would slowly emerge and expand, today comprising a large, decentralized network of agencies, units, and programs. Thirteen of these agencies are known as principal federal statistical agencies, the largest of which by budget are typically the Census Bureau, Bureau of Labor Statistics, National Center for Education Statistics, and National Agricultural Statistics Service. By statute, the director of the Office of Management and Budget (OMB) coordinates this system, but the chief statistician of the United States carries out its functions to ensure the efficiency, effectiveness, and integrity of information collected for statistical purposes.

What and Why to Advocate?

In the 21st century, statistical science is a broad and expanding profession that continues to be guided by an implicit contract with the public. An objective, trustworthy, and independent data infrastructure helps our discipline fulfill this obligation to serve the public good, but its mere existence does not meet this burden. For data to be used for the public good, statisticians should advocate for this infrastructure to help address its new and longstanding challenges.

Consider the surprise 2018 geographic relocation of the Economic Research Service from Washington, DC, to Kansas City, Missouri, which badly damaged the agency’s workforce and led to a robust debate about the purported political reasons for such a change in locale. Or the administration’s effort to add a citizenship question to the ongoing 2020 census, despite a disapproving outcry from myriad scientists and citizens who feared degradation of quality and potential politicization of the nation’s only constitutionally mandated survey. The Supreme Court ruled the stated reason to include such a question was “contrived,” but widespread controversy arguably damaged the data infrastructure, which endeavors to be apolitical and dictated by science.

Even today, efforts remain to compile data on citizenship using information initially collected for regulatory or other administrative purposes, further fraying the public’s trust in government statistical agencies.

These and other events make clear the need for scientists to advocate for the integrity of statistical agencies so they operate according to best practices, such as those defined by the National Academy of Sciences (NAS) and OMB.

The National Academies’ Principles and Practices for a Federal Statistical Agency (6th ed.) and OMB’s Statistical Policy Directive #1 offer insight about the values statisticians’ advocacy efforts should emphasize. Trustworthiness, objectivity, nonpartisanship, relevance, and timeliness, to name just a few, should drive our data-related advocacy efforts, as should principles summarized in the United Nations Statistics Division’s Fundamental Principles of Official Statistics.

The ASA’s own Ethical Guidelines for Statistical Practice describe a collective “responsibility to the public,” and I think good stewardship of this commitment is reflected in advocacy that comports with these values. When should such advocacy occur? What exactly does advocacy look like in practice?

When and How to Advocate?

In April 2017, the ASA Board of Directors discussed criteria to inform when the association should make an official statement. After a thoughtful and prolonged deliberation, the board identified the following five criteria:

• Does the issue impact the statistics profession or the ASA itself?
• Does this issue impact the quality or integrity of science or the ability to inform public policy?
• Does the issue provide an opportunity to educate the public about statistics or the statistical aspects involved in this issue?
• In our judgment, would our members expect us to be involved in the issue?
• Does the issue relate to possible infringement on scientific freedom or human/civil rights?

These questions are by no means exhaustive, but they provide guideposts for when the ASA should speak out on some issue. In addition, these questions can help inform individual statisticians about when to speak out so data is used to advance the public welfare.

To advocate for our data infrastructure, statisticians can get involved in professional associations including the ASA, such as through volunteering with a committee, participating in relevant sections, or joining with student affiliate groups. Additionally, statisticians can join or follow organizations like the Council of Professional Associations on Federal Statistics, the Association of Public Data Users, and the Data Coalition to receive updates about federal statistics via listservs, webinars, or policy briefs.

Signing up for updates about implementation of the Federal Data Strategy, which is the government’s ongoing effort to leverage data as a strategic asset, is another important way to stay engaged with changes to our data infrastructure.

The ASA’s own Count on Stats public relations campaign, which aims to increase public trust in federal statistical agencies, also provides regular information through its newsletter, members only LinkedIn group, and Twitter feeds.

Statisticians who are interested in promoting the integrity of official statistics should also follow the International Statistical Institute and its global network of associations working to apply “statistical science for a better world.”

Professional associations also sponsor advocacy opportunities, such as the ASA’s recent member-driven initiative about staffing and budget problems at the National Center for Education Statistics.

On an individual basis, constituents can request to meet with their elected official to advocate support for statistical agencies and highlight society’s collective need for the objective, nonpartisan data collection and analysis services statisticians provide. Statisticians who meet or communicate with their public representatives should educate themselves about proposed policy changes that have widespread support by the broader scientific community, such as those highlighted by the bipartisan Commission on Evidence-Based Policymaking.

Another way to get involved is by using statistical skills to analyze proposed regulations relevant to statistical science, which are frequently published in the Federal Register, the government’s official vehicle for seeking public comment. As an example, the ASA’s Committee on Scientific and Public Affairs Advisory Committee recently responded to a Federal Register notice about proposals to alter the nation’s poverty measure, developed by the statistician and economist Mollie Orshansky.

Advocacy efforts are enhanced through coalition building, so statisticians should ideally collaborate with a diverse set of colleagues to ensure messaging is clear, consistent, and aligned with best practices. Each of these approaches offers ways ASA members can advocate for our data infrastructure to promote the use of data for the public good.

What You Can Do Now

To most effectively use data for the public good, society needs a strong, trustworthy, and objective federal data infrastructure. By engaging in ethical and nonpartisan advocacy, statisticians can demonstrate shared commitment to our social contract, which dates to at least the mid-19th century in the United States.

Leaders from our data infrastructure, including Katherine Wallman, former chief statistician of the United States; Charlie Rothwell, past director of the National Center for Health Statistics; and Steve Landefeld, previous director of the Bureau of Economic Analysis, have described many of the challenges that need to be addressed through advocacy and, ultimately, policy change. Among others, these include instances of political interference, depreciating top-line budgets, diminished autonomy and authority, stagnant hiring practices, aging technology and equipment, and technical problems like declining survey response rates.

Expert statisticians across federal and state governments are to be commended for continuing to deliver the high-quality data collection, analysis, and reporting services we need, but their success is driven by skill and a personal commitment to use data for the public good. The broader statistical community must continue to champion their essential work and our need for a federal statistical system that reflects best practices. Advocating for our federal data infrastructure offers one way to help ensure that publicly financed data collection, analysis, and reporting activities are positioned to increase the role of evidence-based policymaking.

Further Reading
44 U.S.C. § 3504(e)(7)

American Economic Association Committee on Economic Statistics. Opportunities to comment on proposed federal data collections: A primer on how to respond.

Anderson, Margo. 2016. ASA and census … A long and productive relationship. Amstat News.

ASA Ethical Guidelines for Statistical Practice.

Citro, Constance F. 2016. The US federal statistical system’s past, present, and future. Annual Review of Statistics and Its Application 3(1):347–373.

Groves, Robert M. 2010. The structure and activities of the US federal statistical system: History and recurrent challenges. Annals of the American Academy of Political and Social Science 631(1):163–178.

Landefeld, Steve. 2020. State of the US data infrastructure at the Bureau of Economic Analysis.

National Academies of Sciences, Engineering, and Medicine. 2017. Principles and practices for a federal statistical agency: Sixth edition. Washington, DC: The National Academies Press.

Regan, Opal G. 1973. Statistical reforms accelerated by sixth census errors. Journal of the American Statistical Association 68(343):540–546.

Rothwell, Charles. 2019. State of the US data infrastructure at the National Center for Health Statistics.

Wallman, Katherine. 2019. State of the US data infrastructure.

United Nations Statistics Division Fundamental Principles of Official Statistics.