This column is written to inform ASA members about what the ASA is doing to promote the inclusion of statistics in policymaking and the funding of statistics research. To suggest science policy topics for the ASA to address, contact ASA Director of Science Policy Steve Pierson at pierson@amstat.org.

Column Editor Steve Pierson earned his PhD in physics from the University of Minnesota. He spent eight years in the physics department of Worcester Polytechnic Institute and later became head of government relations at the American Physical Society before joining the ASA as director of science policy.

The ASA’s science policy activity has covered a variety of topics since the creation of the ASA’s science policy staff position 12 years ago. In her September 2019 column, 2019 ASA President Karen Kafadar mentioned several recent ASA science policy and advocacy initiatives, including Count on Stats, protecting a USDA statistical agency, and ensuring the integrity of the decennial census. She also touched upon the ASA’s forensic science reform work, in which she has played a leadership role over the last decade.

The responsibilities of the position are to raise the profile of statistics and statisticians in policymaking and to advocate for the interests of statisticians. The execution of these responsibilities can be grouped into the following non-exclusive categories:

• Statistics improving governance, justice, democracy, and other aspects of society
• Scientific freedom and human rights
• Scientific integrity
• Science to inform policymaking
• Evidence-based policymaking
• Improving science and its process
• Nominations

As you read this, it should be clear that the following activities extend beyond the science policy staff, which now includes a science policy fellow. This work has been accomplished with the help of or by members, committees, sections, ad hoc groups, and task forces, sometimes without the input of the ASA science policy staff at all.

Statistics Improving Governance, Justice, Democracy, and Other Aspects of Our Society

Statistics has the potential to improve broad aspects of our society and every-day life. As part of the responsibility to raise the profile of statistics and statisticians in policymaking, we have supported the following activities that do just that:

Election Integrity

The controversies clouding the 2000 US presidential election led the ASA Scientific and Public Affairs Advisory Committee (SPAAC) and various ASA members to investigate ways statistics can help bolster election integrity. Their work led to various advances, as well as partnerships with many organizations working in the area. One such partnership brought about the ASA playing an integral role in the development of the 2008 Principles and Best Practices for Post-Election Audits. That document, which the ASA endorsed, encouraged risk-limiting audits (RLAs) to take the place of auditing a fixed percentage of ballots, no matter the margin and with no scientific justification for the percentage.

Philip Stark laid out the framework for rigorous RLAs in 2008 and went on pilot them with various California counties. In 2010, the ASA explicitly endorsed RLAs, recommending they be routinely conducted and reported in all federal and most state-wide election contests.

Though activity by the ASA on election integrity has slowed in recent years, Stark remains actively involved in advancing RLA theory and methods—the latest approaches relying heavily on sequential tests derived from martingale inequalities—in addition to achieving their wider use.

RLAs have been piloted in at least nine US states and Denmark. RLAs are required by or mandated in statute in California, Colorado, Nevada, Rhode Island, Virginia, and Washington. Further, earlier this Congress, Sen. Ron Wyden introduced a bill requiring RLA’s, which the ASA endorsed.

Forensic Science

Kafadar reviewed in her September 2019 article the ASA’s forensic science work—under the guidance of the ad-hoc Advisory Committee on Statistics in Forensic Science—noting board statements on the importance of statistical research in strengthening forensic science and recommendations for the use of statistical statements in expressing the strength of forensic evidence. She also mentioned the National Institute for Standards and Technology (NIST) funding of the Center for Statistics and Applications in Forensic Evidence (CSAFE). There are other indications of the central role of statisticians in forensic science reform, including Kafadar’s September 2019 congressional testimony, Constantine Gatsonis’s December 2011 testimony, the strong engagement of the NIST Organization of Scientific Area Committees for Forensic Science with statisticians, and the CHANCE special issue on forensic science.

Use of Value-Added Models for Evaluation of Teachers

In 2014, the ASA board issued a position statement to better inform the use of value-added models (VAMs) for educational assessment. With technical input and guidance from Sharon Lohr, Daniel McCaffrey, and Walter Stroup, the statement noted the strengths and limitations of VAMs and made recommendations for their use. While it’s hard to measure the impact of the statement, it has been cited extensively in discussions about the use of VAMs for the evaluation of teachers.

Statistical Perspective for Federal Calls for Comment

The ASA, through its committees, has been active in responding to federal calls for comment. For example, SPAAC has provided statistical perspective on the policy comment embargo time for federal economic statistics, the federal poverty measure, a citizenship question on the decennial census questionnaire, and the proposed EPA transparency rule (as noted below) in the past two years. The ASA Privacy and Confidentiality Committee (P&CC) has also been instrumental in the ASA’s responses to calls for comment.

Scientific Freedom and Human Rights

Through the leadership of the ASA Scientific Freedom and Human Rights Committee, the ASA monitors these issues as they relate to the statistics profession. The principal activity in this regard over the past decade has been advocating for government statisticians around the world who have been persecuted for their professional work.

While there are numerous examples, the most prominent have been Graciela Bevacqua of Argentina and Andreas Georgiou of Greece. This work has included letters, petitions, meetings, media outreach, and a board statement. We posted a policy update on the ongoing Georgiou case to the ASA website in September.

Scientific Integrity

Scientific integrity is a strong running theme in the ASA’s science policy work. Indeed, the ASA’s positions on the late addition of the citizenship question to the 2020 Decennial Census questionnaire and the relocation of the USDA Economic Research Service (ERS) are based on maintaining the integrity of both agencies’ products.

The science that should inform regulations is another example of this work. Under the leadership of SPAAC and P&CC, the ASA has been vocal about proposals to base US Environmental Protection Agency (EPA) regulations on research for which the underlying data is publicly accessible, starting with House-passed bills known as the Secret Science Act and HONEST Act.

While the ASA applauded the idea that researchers and federal agencies strive to make data available to others—under strict pledges to maintain confidentiality of data provided by individuals and establishments where necessary—and to encourage reproducible research, the ASA urged a 2018 proposed rule from the EPA not be adopted because, among other reasons, it hampers the use of evidence, introduces potential bias in rulemaking, and neglects the challenge and complications of protecting confidentiality. In 2017, then ASA President Barry Nussbaum and SPAAC Chair Jerry Reiter published an op-ed in The Hill, “HONEST Act Needs Honest Engagement of Scientific Community.”

Science to Inform Policy

Scientists and scientific associations share a common desire for the science used to inform policymaking to be robust and represent a balanced view. Recognizing their expertise ends at the science, scientific associations generally stop short of recommending action or endorsing a specific policy approach. The ASA’s climate change work fits well within this category of encouraging policymakers to use sound science to inform climate-related discussions.

Soon after the hiring of the ASA science policy director, an ASA member commented to the current executive director in 2008 that she didn’t think the ASA was addressing the country’s major challenges, with climate change being one. That conversation led to a series of actions by the ASA overseen by the since-created ASA Advisory Committee on Climate Change Policy and its precursor. For example, the committee organized a special issue of CHANCE, provided a compelling case for a federally funded interdisciplinary research program around climate, and wrote an Amstat News article about the status of climate change science that was viewed more than 23,000 times.

The 2008 conversation also led the creation of an inter-societal Climate Science Working Group (CSWG), an informal assembly of Washington representatives who meet monthly to share information related to climate discussions in the federal government and coordinate activities. The coordinated activities include the annual Climate Science Day, for which a few dozen scientists of various disciplines form teams of two to three individuals and—through meetings on Capitol Hill—seek to establish connections with congressional lawmakers and their staffs on climate science.

Another CSWG activity was the 2016 letter to Congress signed by 31 scientific societies sharing a consensus view of climate change, which is—in short—that climate change is occurring and greenhouse gases emitted by human activities are the primary driver. Associations in the CSWG include AAAS, the American Geophysical Union, the American Chemical Society, and the American Meteorological Society, as well as three agriculture-related associations.

Evidence-Based Policymaking

Evidence-based policymaking is another strong theme throughout the ASA’s science policy work, because it represents a general desire of the ASA membership to ensure policymakers consider the available evidence in their work. The statistical community is also supportive of the closely related data-driven decision-making.

The ASA’s support of the effectiveness and efficiency of the federal statistical agencies could be its own heading because of the ASA’s history of supporting the agencies going back to its 1839 founding. I categorize this support under evidence-based policymaking because, as the ASA Board stated in December 2018, “The federal statistical system forms the foundation of US evidence-based policymaking and data-driven decision-making.” The board endorsement of the National Academies’ Principles and Practices for a Federal Statistical Agency provides important parameters for the ASA’s work.

The ASA’s work in support of the federal statistical agencies has several components: (i) ensuring the integrity of their data and products, as noted above; (ii) maintaining objective, credible, and timely data through sufficient agency independence via control over their publications, budget, and information technology; (iii) providing sufficient resources to do their jobs effectively and efficiently; and (iv) building trust and confidence in government statistics.

Beyond census citizenship and ERS work mentioned under Scientific Integrity and the ASA’s support of the federal statistical agencies’ budgets and tracking thereof, the ASA evidence-based policymaking work is represented by the following activities:

Count on Stats

Count on Stats (CoS) is a public relations campaign in its third year meant to build public confidence and trust in government statistics. Through CoS, we have advocated for a rigorous 2020 census, highlighted the value and necessity of the ERS, distributed weekly social media features, and provided general agency support though congressional outreach by ASA staff. This advocacy contributed to outcomes such heightened media attention on statistical agencies, including publications in The Washington Post, The Des Moines Register, and The Kansas City Star.

In spring 2019, we launched the CoS LinkedIn group, which is freely available to any person with a LinkedIn account and serves as a networking and resource group for anyone with an interest in federal statistics. One prominent feature of the CoS LinkedIn group is the ongoing State of the US Data Infrastructure series, which is a monthly campaign to assess the challenges facing principal federal statistical agencies. Katherine Wallman wrote the inaugural post, “The State of the US Data Infrastructure.” Former National Center for Health Statistics (NCHS) Director Charlie Rothwell assessed the state of NCHS infrastructure, and former ERS Administrator Katherine Smith Evans wrote about the state of ERS infrastructure. More are underway.

Independence of the Federal Statistical Agencies

The ASA has been an active advocate for the independence of the statistical agencies so their data is widely viewed as objective, in accordance with Principles and Practices for a Federal Statistical Agency and the White House Office of Management and Budget (OMB) Statistical Policy Directive No. 1: Fundamental Responsibilities of Federal Statistical Agencies and Recognized Statistical Units. This work has ranged from helping to ensure the IRS Statistics of Income (SOI) Division maintained control of their IT and helping to bolster the control of the Bureau of Transportation Statistics over its publications, IT, and budget to keeping the ERS in the USDA research mission and advocating to restore Senate confirmation of the heads of the Bureau of Justice Statistics and National Center for Education Statistics.

Puerto Rico Institute of Statistics (PRIS)

The ASA was an active partner in advocating for the PRIS in 2018. In addition to disconcerting meddling by the governor on the PRIS advisory board, there was a proposal to consolidate it into another government department, remove its independence protections, and outsource its statistical research. PRIS was saved in a last-minute move by the Puerto Rico legislature, but concerns remain.

Commission on Evidence-Based Policymaking (CEP)

The ASA has been supportive of the CEP from its beginning, including board endorsement of the CEP recommendations, and remains so through enactments of its recommendations.

Improving Science and the Scientific Process

Statisticians firmly believe more engagement of statisticians throughout the research process will improve both the science and scientific progress. The ASA’s work here has largely been through the federal research funding agencies that help set research funding policy and carry it out. One component of this work is supporting the budgets for a few federal research funding agencies.

Highlighting the Role of Statistics

The ASA has been active in highlighting how more engagement of statisticians and statistics could improve the science. In a 2012 guest editorial in Science advocated by the ASA, Marie Davidian and Thomas Louis answered the question posed in the title of their piece, “Why Statistics?”

In 2014, three groups made the case for how statistics could contribute to three presidential initiatives. The three whitepapers were the following: “Discovery with Data: Leveraging Statistics with Computer Science to Transform Science and Society,” “Statistical Research and Training Under the BRAIN Initiative,” and “Statistical Science: Contributions to the Administration’s Research Priority on Climate Change.” As an example of the impact of these white papers, there were reports that the BRAIN whitepaper was circulated widely at National Institutes of Health (NIH) and one of the BRAIN funding opportunity announcements (FOA) noted the whitepaper as having inspired ideas in the FOA.

In 2015, the board issued a Statement on the Role of Statistics in Data Science, saying statistics—along with computing and database management—is foundational to data science. The aim of the statement, which has been viewed nearly 15,000 times, was to help “reinforce the relationship of statistics to data science and further foster mutually collaborative relationships among all key contributors in data science.” As one part of the broader effort in this regard, Amstat News has been running Q&As with representatives of new master’s and doctoral programs in data science and analytics to “highlight the programs that are cross-disciplinary and engage statisticians.” Collectively, the pieces have been viewed 45,000 times.

Statistics Improving the Science

The ASA Committee on Funded Research (CFR) has been active in highlighting the role statisticians and statistics can play, especially in improving federally funded research. In 2017, they issued two guides, one for nonstatisticians to improve the science in their funding proposals and one for statisticians to serve effectively on funding review panels. The CFR also urged a meeting of an ASA group with the leadership of the NIH Center for Scientific Review (CSR), which oversees the consideration of proposals for more than $20 billion in NIH extramural research funding. That meeting led to a 2016 article in the CSR publication Peer Review Notes, “Statisticians Share Insights for Applicants and Reviewers.”

Another ASA visit to NIH to discuss data science, led by former ASA president Marie Davidian—who also led the CSR visit—eventually led to a group headed by Nancy Reid publishing a piece in PLOS Computational Biology, “Ten Simple Rules for Effective Statistical Practice,” that has been viewed more than 235,000 times.

P-Values

While the ASA science policy office has not played a role in the ASA’s recent p-value work, I include it as likely the ASA’s most influential work in bettering science. The messages of the 2016 statement—viewed more than 360,000 times and cited more than 2,400 times as of November 2019—quickly rippled throughout the scientific community, sparking wide discussion about the use of the p-value. These discussions escalated to research entities, federal agencies, and journals (including The New England Journal of Medicine).

Reproducibility and Rigor of Science

The rising profile for statistics over the last 10 years has been helped by the rise of big data, data science, and the concerns for reproducibility in science. Statisticians have played a lead role in these discussions, through the p-value discussions and more broadly. In early 2017, an ASA-organized group provided recommendations to funding agencies for supporting reproducible research.

Nominations for Federal Committees/Boards, Awards, and Other Honors

The ASA is active in supporting nominations to federal statistical advisory committees and boards because of the valuable role statisticians can play on such committees in informing policy and because it helps raise the profile of the profession. The ASA’s nominations more broadly have been enhanced through the creation of the External Nominations and Awards Committee.

Advocating for the Profession: The ASA Wants Your Recommendations

What issues would you identify as opportunities for ASA science policy work? Are there policy issues in which more statistical perspective or input would benefit the process and/or product? What’s happening in your community (geographical or otherwise) about which the ASA’s input could be instrumental? How can the ASA raise the profile of statisticians to inform evidence-based policymaking?

If you have responses to any of these questions, the ASA wants to hear from you! Through this column, I wish to emphasize the opportunity of—indeed, the importance of—ASA members providing input and engaging in ASA science policy work.

Advocacy for the ASA means supporting the profession of statistics, encouraging the engagement of statisticians, providing the perspectives of the ASA as it relates to issues on which we have expertise, or urging a particular course as it relates to the ASA perspective.

To recommend a topic for ASA advocacy, you can email the ASA director of science policy. Alternatively, you can contact the ASA executive director, a board member, the ASA Scientific and Public Affairs Committee (SPAAC), or another relevant ASA committee with your request. After receiving a request, the ASA will respond through the following actions:

1. Assess whether the topic meets the criteria provided by the board in 2017.
2. Consider the likelihood of success, which includes asking whether we can leverage concern in our broader community.
3. If a topic meets the board criteria and appears to be a relevant issue the ASA can influence, we assess whether it is in accordance with a board statement or previously board-approved action or the board must be consulted on the specific issue.

The ensuing actions the ASA could take include responding to calls for comments, writing a letter, issuing a board statement, requesting more information, meeting with key officials, and issuing a whitepaper.

In addition to SPAAC, other ASA committees and sections help ensure ASA science policy and advocacy work engages broad and relevant expertise across the association. For health policy issues, the ASA Health Policy Statistics Section is the group to be consulted, whereas issues relating to climate change are relevant to the ASA Advisory Committee on Climate Change Policy (ACCCP). Other frequently consulted committees include the Advisory Committee on Forensic Science, Privacy and Confidentiality Committee, Committee on Funded Research, and Scientific Freedom and Human Rights Committee.

Besides suggesting topics for ASA science policy work, you can also get involved by serving on a committee, raising ASA advocacy issues with your elected officials, or being a science policy fellow with the ASA or through the AAAS program. You could also seek out a federal advisory committee matching your expertise and ask to be nominated by the ASA.

To keep abreast of ASA science policy work, you can read the science policy section of the Member News—emailed monthly to ASA members—monitor the news and highlights section of the ASA science and policy homepage, follow @ASA SciPol on Twitter, connect with me on LinkedIn, and watch for updates in Amstat News.