Introduction by ASA Director of Science Policy Steve Pierson

For the latest State of the Data Infrastructure (SDI) feature, Count on Stats spoke with three experts about the US Census Bureau: Margo Anderson of the University of Wisconsin, Milwaukee, an expert on the history of the bureau and federal statistical system; Allison Plyer, chief demographer of The Data Center and recent chair of the bureau’s Scientific Advisory Committee; and Nancy Potok, most recently chief statistician of the US (2017–2019) who, most notably, has also been deputy director and chief operating officer of the bureau and deputy undersecretary for economic affairs at the US Department of Commerce.

The interview took place in early December, with the bulk of the 2020 Census work wrapped up and a month prior to the swearing in of Robert Santos as bureau director. After a decennial census challenged by the pandemic, concerns of political meddling, extreme weather and wildfires, and complaints over its new disclosure avoidance system, the bureau must assess the lessons learned and anticipate unknown challenges to be faced in 2030 while also addressing challenges such as declining response rates and increased demand for more timely and granular data. Anderson, Plyer, and Potok discuss these challenges; the bureau’s vulnerabilities exposed in the last several years; and the support needed from Congress, the administration, and data users.

Having witnessed closely the 2020 and other census operations through a pandemic and concerns of outside interference, what stands out to you about the capabilities, challenges, and vulnerabilities for the US Census Bureau?

Nancy Potok is the chief operating officer of NAPx Consulting. She served as the chief statistician of the United States in the executive office of the president until January 2020, during which time she also served as a commissioner on the US Commission on Evidence-Based Policy Making and co-chair of the Federal Data Strategy.

Nancy Potok: I have great admiration for the tremendous work of the Census Bureau professional staff through the pandemic, including on the decennial census, and for much of the behind-the-scenes work that took place to keep all the census operations going that is largely unknown to the public. Many congratulations to the staff for that. Because I care a lot about the Census Bureau, I’m going to first share what I see as a major vulnerability in the interest of perhaps helping the bureau and whole federal statistical system become even better.

Here is an illustration of that vulnerability: As the chief statistician of the US in the Office of Management and Budget (OMB) when we were considering adding the citizenship question, many looked to me as the stopgap for the question’s inclusion on the 2020 Census. That’s because the Paperwork Reduction Act (PRA) gives authority to OMB to approve the content of federal statistical data collections, which should ensure the quality and objectivity of federal statistical programs. The PRA, however, actually provides that authority to the director of OMB, who generally has delegated that to the chief statistician through the politically appointed administrator of the Office of Information and Regulatory Affairs in OMB. I have no doubt that had the Supreme Court decision gone the other way and I had disapproved the census questionnaire containing the citizenship question, that decision would have been overridden by the director of OMB.

In addition, the secretary of commerce proposed the citizenship question to Congress without notifying OMB, at least through the normal channels by which all the other questions had been sent forward. To protect the integrity of the statistical system from political interference in the future, I strongly recommend strengthening the authority of the chief statistician and modernizing various provisions of the Census Bureau’s authorizing statute, Title 13 (i.e., the portion of the US Code on the operations, responsibilities, and authorities of the US Census Bureau).

On the modernization topic, the Census Bureau has done a great job in setting up the household and business pulse surveys and the Frames Program. Both illustrate what the Census Bureau can do and where the Census Bureau needs to be headed from a data standpoint.

One of the strengths of the Census Bureau is its ability to build collaborations with the other statistical agencies. The leaders of the bureau’s demographic and economic programs were able to set up and conduct the pulse surveys in partnership with multiple agencies, even when the entire staff was remote due to the pandemic. OMB moved at a rapid speed to approve the surveys. Many of the questions in those surveys came from the other statistical agencies and addressed important questions about the state of the nation during the pandemic, a rapidly changing environment.

I’d like to see more support for the bureau to be able to react quickly and do more of those types of things that inform important policy decisions in a timely way. We shouldn’t have to be in a crisis situation for the Census Bureau to get the resources and support it needs to be agile and responsive in areas beyond the decennial censuses.

One of the biggest challenges for the Census Bureau is the need to accelerate its efforts to build a culture that’s much more responsive to its users. The bureau needs to be nimbler in terms of its ability to maintain quality of data but still respond to what users want and expect from data and information providers in today’s world. It has to move faster to adopt innovative approaches at a pace that keeps up with data innovations outside of government. People depend on census data, but they need it faster and at a more granular level without the quality being compromised.

I think there are several steps the Census Bureau can take to accomplish that. During the last century, the Census Bureau led the way in changing federal statistics; it’s time for another major leap forward. For this, the bureau will need a lot of support from Congress and the administration and more partnerships with academia and the private sector, as well as substantial input from the users.

Allison Plyer is chief demographer and coauthor of Pandemic to Prosperity, developed in collaboration with the National Conference on Citizenship to track the impact and ongoing recovery from the COVID health and economic crises. Plyer is also coauthor of The New Orleans Prosperity Index, which examines the extent to which economic outcomes have improved for Black New Orleanians since the end of the Civil Rights era.

Allison Plyer: Let me first commend the bureau. They prioritized science when it came to the 2020 Census data. They did not release data before they felt it was ready and resisted political pressures to do so. I think we should all be relieved and grateful to see they still represent that bulwark.

I also want to echo Nancy’s point for providing data faster. We need to realize COVID should be categorized as a disaster. It caused an enormous break in the status quo, followed by dramatic changes in consumer behavior, and now major policy changes for recovery. That means all the things we measure are changing much, much faster. The one-year or more lag in data makes it nonrelevant by the time it’s released. We can expect this kind of rapid change for several more years, if not forever. We need rapid statistics constantly now. Consider, for example, monthly poverty estimates. Just like we need good data on jobs being created, we need good timely data on whether government programs are having an impact on poverty. To have a sense of what the current policies are doing for Americans, we need more timely estimates of poverty and income.

For challenges, I think of two things. First is the Census Bureau faced unprecedented challenges with the decennial census, including political interference, the pandemic, and historic weather events.

These challenges raised new concerns about data quality. At the Census Scientific Advisory Committee, we called for radical transparency from the Bureau to keep the country’s trust in the face of these unprecedented challenges.

To their immense credit, they have provided a lot of that. Now that the apportionment and redistricting data have been released, we really need to use all the avenues available to improve the data that’s going to inform federal funding distribution for the next nine years.

The challenge is that it’s the jurisdictions with poorer quality 2020 Census data that need to provide local data to improve the annual population estimates, but neither the Population Estimates Challenge Program, the Count Question Resolution Program, nor Special Censuses have had a lot of return on investment in recent years.

Before 2010, we were able to increase New Orleans’s population estimate by 75,000 people, correcting estimates that would have otherwise been 25 percent too low. Post 2020, we need an innovative new approach to the Population Estimates Challenge Program that blends several local data sets and methodologies to improve the accuracy of state and local estimates.

The second challenge is declining survey responses. The 2020 American Community Survey (ACS) one-year failure to meet statistical standards is a red flag after decades of declining survey response, and this is going to be an ongoing issue.

The pandemic hindered ACS operations, but the differential response rate by income also widened. Because disasters accelerate preexisting trends, overall survey responses may continue to drop at an accelerated rate going forward, and the likelihood of lower-income households and other hard-to-count populations responding may remain even more diminished.

And it’s a broader epistemic challenge for our society. We all know how unreliable political polling has become. The ACS has a sample size greater than 2 million and a big notice on the outside that response is mandatory by law. It is the very last survey that ought to be affected by declining survey response, but it has been. This should concern anyone who wants to know about our people, places, and the economy, and it’s why we really need to invest in the bureau.

Margo Anderson, distinguished professor emerita from the University of Wisconsin, Milwaukee, specializes in urban history, particularly of Milwaukee, and the social and demographic history of the United States. She has authored several publications, including the second edition of The American Census: A Social History.

Margo Anderson: The Amstat News data infrastructure series should be seen perhaps as just the most recent step in the American tradition of using statistical information to address the public needs of society. We can thank the framers of the Constitution for their insight into the importance of data for decision-making. They addressed the problem of allocating political representation (and originally tax responsibilities) among the states on the basis of a decennial census. In my view, these men embedded the “data infrastructure” in the very foundation of the American state. And they did so before there was much of a statistical profession or rules and procedures of how to do it. So, given how much is at stake, it’s always been a challenge to take the census fairly and efficiently.

Tying into Nancy’s point on engaging data users and Allison’s point on survey response, the Census Bureau, the American public in their roles as respondents and data users, and politicians and data scientists all engage in a complicated conversation of how to collect information. We have to come to consensus on what to do, who is supposed to do what, and whether the resulting information is good “evidence.”

These questions are pretty stark at the moment because the 2020 Census faced unprecedented problems, particularly political meddling with the operational plans for the count by the Trump administration and the pandemic disruption of field administration. We’re literally working out answers now about how to use the 2020 Census results for the next decade.

Please elaborate on the support needed from the administration to meet these challenges.

Nancy Potok: Making the household and business pulse surveys a reality so rapidly illustrates the kind of support the bureau needs to provide timely data. OMB worked really quickly to approve the surveys, making an exception so the bureau didn’t have to go out for months for public comment in the Federal Register, as is the standard procedure. Money and people also had to be redirected for these surveys to be conducted. However, support for reprioritizing the existing resources for something that seemed to be most critical meant other projects weren’t being done.

And what is most harmful to the bureau is the lack of attention it gets between decennial censuses, as if substantial change doesn’t need substantial investment. I have not seen innovation take place in the private sector without investment. Why is government supposed to innovate out of thin air?

In the end, lack of timely data hurts the entire nation, so not investing in this is really shortsighted. And many of the innovations would actually save money in the end because getting people to respond to surveys is a lot more expensive than using alternative data sources such as administrative records the government has already spent the money to collect.

Census also needs to have the ability to move more quickly with outside partnerships that would bring innovations. When I was at Census, Congress granted the bureau new authority to enter into cooperative agreements. The beauty of a cooperative agreement is the ability to move forward in a research partnership without the heavy lift and longtime horizon on the competitive contracting side.

There’s also a joint venture authority, which allows you to work with outside private sector or nonprofit partners to do research in the areas Census desperately needs. Because Census is part of the Commerce Department, instead of having its own joint venture authority, it has to go through another bureau in the Commerce Department. That adds layers of red tape that don’t need to exist and slow things down.

The same thing with cooperative agreements, there’s another bureau Census is required to go through to enter into cooperative agreements, which slows down the process, as well.

These existing authorities can help the Census Bureau react more quickly to the needs of the country for data, but bureaucracy weighs it down. There is a tension between the Commerce Department trying to be more efficient through shared services and the reality of it layering more bureaucracy and actually reducing productivity. What looks good on paper doesn’t always pan out in the real world.

I think more autonomy for the Census Bureau to be able to enter directly into these partnerships and move quickly would help with that nimbleness and would probably have a positive benefit-cost ratio for bringing innovation into production.

Margo Anderson: From a historical perspective and lessons to be learned, the interesting thing about the pulse survey is it spoke to an immediate need and the planners were able to deploy a trusted methodology and marshal resources quickly to get the survey off the ground. It also, as far as I can see, didn’t generate political opposition. If you look at surveys that have had trouble getting off the ground, you see one of those factors missing, or slowing the innovations.

So, in the 1930s, senior officials in both the Hoover and Roosevelt administrations were not particularly keen on finding out how high the unemployment rate was. What these officials wanted was to get unemployment down, so proponents of the survey had to figure out how to convince them they had a viable methodology and that monthly measurement of employment and unemployment using sampling would provide evidence for workforce policy. That seems like common sense to us now, but wasn’t in the 1930s.

Allison Plyer: I think that with a little encouragement and investment, the Census Bureau and federal statistical system more broadly could do things that would improve data quality and equity, which would help this and future administrations achieve their goals.

Reporting monthly on median household income and the poverty rate seems like a necessary scorecard if you are trying to raise incomes for working people, which is or should be a bipartisan goal. And data shows concern about poverty can’t possibly skew politically as much as some think. As Indi Dutta-Gupta of Georgetown testified to Congress last year, between the ages of 25 and 60, three out of five people in the United States will experience at least one year in the bottom fifth of the income distribution. That’s a lot of voters.

What is needed from Congress, particularly as it relates to the vulnerabilities around outside meddling mentioned previously, including the three additional political appointments in 2019 and 2020?

Nancy Potok: Strengthening the role of the chief statistician is critical. The chief statistician has very little authority. As I mentioned, most of the statutory authority under the PRA goes to the director of OMB, a highly political position, which is then delegated to the administrator of the Office of Information and Regulatory Affairs, another highly political position, and then trickles down to the chief statistician, leaving little authority except for routine things. The chief statistician position should be built up to give that person authority to make decisions that affect the integrity of the nation’s statistical data. Admittedly, given the structure of OMB, that would be challenging to accomplish with a minor tweak.

The inherent weakness of the chief statistician position compares negatively to other countries where the chief statistician has much more authority and independence, such as in Australia and Canada and throughout Europe. We are embarrassingly behind in that respect and should address it. After all, trillions of dollars of foreign and domestic investment in our economy are resting on people trusting the information coming out of federal statistical agencies. If there is a widespread and worldwide perception that the data has been politicized to suit whatever administration is in power, it will have substantial economic repercussions. And clearly, the current safeguards are insufficient.

We need the federal statistical system to be regarded in the same way as the Congressional Research Service, that is, as a nonpartisan information-providing institution. Revisiting the role of the chief statistician could be helpful in that regard. In addition, if you’re looking across the whole federal statistical system, you could start with looking at the Census Bureau as an example.

Congress needs to do a thorough reexamination of Title 13 and provide clarification on several issues to reflect advances in information technology and changing public perceptions on privacy. Without congressional clarification on the privacy protections in the law and whether they are still relevant in today’s world of information, important decisions are left to a small group of lawyers in the Department of Commerce.

Consider, for example, whether the Census Master Address File really needs to be 100 percent protected by Title 13. Most of that information is readily available online. Not allowing localities access to check whether the official address list used by the Census Bureau is complete because of Title 13 can result in data inaccuracies down the road. It also causes the government to duplicate address listings across different agencies.

Additionally, the role of the secretary of commerce needs to be examined carefully, which goes back to the quality issues already mentioned as a result of political interference. Currently, Title 13 provides the secretary of commerce enormous authority and discretion over these technical decisions, even though those decisions have potentially major political ramifications.

On the topic of additional political appointees embedded in the bureau, it would be nice to be able to prevent that from happening again. But what is the solution? After all, the Census Bureau is an executive branch agency in a Cabinet department and under the authority of the president.

There was proposed legislation back in 2009 or 2010 that tried to give the Census Bureau more autonomy. One provision in that bill would have allowed the Census Bureau to send its budget directly to Congress without going through the Commerce Department or OMB review. That resulted in a veto message from OMB on the whole bill because executive branch agencies need to go through the OMB budget process and the bill would have set a dangerous precedent.

You could try to pass legislation that would limit the number of political appointees at the bureau, but the administration would most likely veto any legislation that restricted the president’s discretion in assigning political appointees to agencies beyond what is statutorily required, such as for an agency head.

In the case of the Census Bureau, after the first legislation was subjected to a veto threat and not passed, subsequent modified legislation was enacted in which the Census Bureau director was given a fixed term and certain statistical competency requirements. That did not stop political interference. But we have seen multiple times in different administrations and across different agencies, it’s hard to have a foolproof method for avoiding political interference in an executive branch agency.

There needs to be a real consensus across party lines that having an independent statistical system is a priority, and, so far, that has not materialized in the form of transformational legislation that is enforced by consistent congressional oversight.

Margo Anderson: I believe it’s important to understand how we’ve arrived at the arrangements.

The census is the only constitutionally authorized activity that grounds the statistical system. Everything else is agency derived. With this context in mind, it is important to understand that when the Census Bureau gets a cold, the whole statistical system is under threat. And it raises the question of why the US has such an intricately structured decentralized statistical system.

In principle, the chief statistician is indeed the only locus in the federal government that has the power to intervene in issues like that seen with the citizenship question. However, to Nancy’s point about the lack of authority for the chief statistician vis à vis that of the commerce secretary, it is a sign of why it would be very useful to revisit the provisions of Title 13 and try to understand why we have competing authorities in the chief statistician and commerce secretary.

In simple terms, the chief statistician position didn’t exist when Congress created a permanent agency for the census in 1902 and placed it in the Department of Commerce and Labor in 1903. The chief statistician position was created almost 40 years later, on the eve of World War II, with the consolidation of activities in the Bureau of the Budget, now OMB.

So, we have these legacies of earlier decisions that need scrutiny, and that, as Nancy said, haven’t been looked at for a really long time. So, I agree on the need for Congress to revisit Title 13 and would urge it be done from the perspective of 21st century statistical infrastructure generally.

That’s a tall order and will, of necessity, require Congress to address these questions in a broader fashion than they’ve done, in my view, since the 1970s. Then, Congress spent almost seven years producing major revisions of Title 13, including PL 94-171 and PL 94-521.

Allison Plyer: This is more Nancy and Margo’s wheelhouse, but I do support strengthening the Census Bureau’s independence. As it pertains to congressional oversight, I will say I’ve been struck by something else—how much of Congress’s discussion of the Census Bureau centers on the per household cost of the decennial. As a data scientist, that is about as irrelevant a statistic as I can imagine.

The Census Bureau’s work underpins our democracy—you literally can’t have a House of Representatives without it, or an Electoral College. It is getting much harder, not easier, to get people to respond to the decennial and to surveys. And it’s getting more expensive. It’s not getting less important. Congress ought to just accept that and pay, full stop.

If the federal government distributes $15 trillion per decade based on census numbers, as Andrew Reamer counts, don’t they want that to go to the right places? And then when you look at even more trillions of market dollars moved by economic indicators produced by the Census Bureau, Bureau of Economic Analysis, and Bureau of Labor Statistics, it just makes good business sense to invest in statistical excellence, to optimize private investment decisions, if for no other reason.

And with global capital so relentlessly mobile, the reputation we have for statistical quality and independence is a strength of our markets compared to say, China’s. To maintain that edge, and to do what Madison suggested and mark the progress of our society, I think Congress ought to be demanding more excellence, innovation, and speed from the bureau and the statistical system and investing to make it possible.

During the 2020 Census process, there were concerns about the new disclosure avoidance system and its reliance on differential privacy. What needs to be done by 2030 or in the shorter run for ACS and other products? What do you think needs to be done around the GIS for the next decade?

Nancy Potok: Fundamental to this discussion is the question, “How good is good enough?” We know statistical data is never perfect. There’s always some range of error in it. It would be useful if there was consensus on the tradeoff between first the privacy and confidentiality risks, meaning the reidentification of individuals and businesses in the data; second, the rapidity at which the data can be collected and disseminated and its related utility, because if it is published too late, it’s not relevant anymore; and third, the accuracy of the data, which also affects its utility and fitness for various purposes. Who’s deciding that right now?

The Census Bureau needs to put together a standing outside group on disclosure avoidance to help them decide the tradeoffs between the utility and accuracy of the data and the risk of reidentification, as well as the speed at which data needs to be disseminated. Those decisions should not be made entirely by a small group of people within the Census Bureau. Establishing an outside stakeholder advisory group providing serious advice on disclosure avoidance would be a huge cultural change for the bureau.

However, data users and representatives of the public should be able to weigh in meaningfully and substantially on something as important as the necessary level of privacy protection in comparison to the utility of the data. Census needs an advisory committee with that kind of punch.

While the bureau attempted to do this with the 2020 Census data, it was a difficult, frustrating process riddled with trial and error. I give the bureau high marks for trying to be as transparent as possible, but there has to be a more institutionalized approach going forward, as the issues are not going away with the 2020 Census.

As society changes, the context for privacy changes, and as technology changes, the infrastructure should be there for the Census Bureau to make decisions that serve the people.

I would also suggest the bureau continue to explore alternatives to differential privacy. I think the bureau knows it is not the only path to providing the public with access to important data. For example, improving secure access to data would help a lot, including, but not limited to, substantially modernizing the federal statistical research data centers and making statistical data more readily available in other venues that don’t require a lengthy and expensive background check just to access data. The whole federal statistical system needs to take on that modernization.

The bureau also should move expeditiously to implement a meaningful tiered access system for census data and be more generous in making data available through secure infrastructure run by organizations other than the Census Bureau, which now happens very sporadically. These actions are not only good practice, they are a fundamental piece of the Evidence Act. And that is just one more reason why Title 13 needs to be revisited, to make sure it is in harmony with the need for solid, evidence-based policymaking.

Margo Anderson: Right now, it is only the Census Bureau. Users are challenging the bureau’s approach. Congress has yet to do so, but I suspect that will come, too.

I’ve been trying to broaden the framing of the issues of disclosure control and talk less about “privacy” and more about “data stewardship.” Stewardship implies the issues of how information, once collected, is coded, curated, saved, published, and explained so it can be used. It allows us to ask, “Who is the steward?” “Who is responsible for doing what?” This should be a much more complex conversation with professionals, users, and respondents. Reluctant respondents instinctively allude to these issues when they express resistance to responding, and in my view, may also be claiming the right to be part of the “stewardship” decision-making as well, for example, to ask what other sources of information can or should be accessed. I see the issue of the decline in survey response as part of a reexamination of data stewardship.

Allison Plyer: As Jeri Green of the Urban League observed in another context, census doesn’t lose people’s trust by admitting to shortcomings; it’s the opposite. The Census Bureau accomplished something difficult: It built an algorithm in public for disclosure avoidance, reflecting public values, and received input from its most severe critics in setting the level of privacy loss. They’ve implemented disclosure avoidance in the past, but never in such a transparent way. The promise of the system they went with for 2020 was a quantifiable and transparent allocation of privacy loss.

This system is not perfect, and there are more difficult decisions ahead about privacy settings for later data products. Going forward, there are other approaches we also ought to be thinking about. For example, the private sector is using synthetic data for investing and planning without hesitation. We probably need a lot of changes in how we develop our census data in general that may make the debate about disclosure avoidance very different in a few years.

Nancy, you were instrumental in the ASA Task Force on the 2020 Census Quality Indicators, not only leading it but calling for such an effort with others in late summer 2020. Looking to the 2030 Census, how can the bureau, Congress, and the administration be prepared to facilitate earlier assessment of the data quality?

Nancy Potok: First, I encourage everyone to read both the October 2020 report and the September 2021 report produced by the task force. Both reports contain several task force recommendations, which include the need to have an earlier assessment of the census quality in 2030.

We recommended that the Census Bureau be forward-looking and open to new approaches, including greater transparency for the operational data at low levels of geography. The bureau collects operational data in real time, such as whether a household has self responded, its information was collected by an in-person interview, or its information had to be included through a proxy interview.

There is often a gradation of quality in the responses, if they are collected by proxy relative to information provided directly by the residents of a household. It is helpful to know whether such indicators are clustered, and this is indeed knowable and should be publicly available before the numbers are released. Additional information, such as where statistical imputation was used to fill in missing data during processing, should also be available. The bureau now has the technology to do this and can plan for it in 2030.

In addition, the task force wanted to see more transparency around the quality of administrative records used and where they were used. Administrative records had a big impact during the 2020 Census, and the bureau needs to start now to make sure relationships are in place to maximize the use of records for 2030. That includes relationships with other government agencies and institutions outside of government, such as universities. But these relationships are built on trust and mutual benefit, so it is important that institutions in particular can see the value of the relationships before they end up potentially undercounted in 2030.

The Census Bureau seemingly feels constrained by Title 13 in terms of releasing some of the operational data, but that constraint is subject to interpretation. One of the recommendations of the task force was to reexamine Title 13 to both increase the independence of the Census Bureau and address these grey areas of what actually is needed to protect confidentiality. Outside the Census Bureau, I’m not sure many people believe this operational data or paradata requires the same level of protection as the actual personal data from a household. I hope Congress takes this up in a reexamination of Title 13.

The so-called gold standard of assessing quality is the post-enumeration survey conducted by the Census Bureau, which is ideally conducted in parallel with field operations. The processing of that survey can be sped up considerably, and the bureau should prioritize work to make it available much sooner. Because that survey was at the heart of some controversy in 2000, its timing was pushed back and size reduced in 2010. That’s not a reason to delay it into perpetuity. The public deserves to understand the quality of the data coming out of the census in a timely way, not years later.

There are some other files the bureau could make available sooner if it would prioritize the processing of these data sets. That would help interpret the demographic analysis and population estimates as quality benchmarks, as well. Census, the Commerce Department, and Congress should make it a priority to provide the resources to get this done.

To close, please summarize the Census Bureau’s health, resources, and state of data infrastructure.

Allison Plyer: The 2020 Census was not the best census ever, but it was no doubt the best count possible under very difficult circumstances. Bureau staff are tired, and no wonder. The 2020 Census didn’t fix the historic undercounts of people of color, the indigenous, or people in rural areas. And with the differential nonresponse in ACS data, our statistics are starting to skew to make Americans look wealthier, whiter, more likely to be married, and more likely to own homes. Compounding the problem is that all these “easy-to-count” people are already in administrative records and the private sector is keeping track of them just fine.

Going forward, we need to figure out how to get good data on the hard-to-count, and we need that data on a more frequent and timelier basis. If we don’t focus on changing how we gather data so we prioritize marginalized groups, we will lose the values added by public statistics: universality and fairness.

Margo Anderson: I’d return again to the fundamental need for good information to ground democratic decision-making. That base understanding sometimes gets lost in the day-to-day processes of appropriations and budget writing. Congress hears agency officials asking for money or pointing out the erosion of budgets over time. They hear agencies touting the promise of their latest scientific innovation if there was just more money. They hear about how much trouble the agency faced taking the 2020 Census.

To focus the resource “ask,” one should nevertheless acknowledge the 2020 Census got done. We reapportioned, we are redistricting, and we’re going to use this information for the next 10 years. So, it’s time to think forward, if you will, and acknowledge that the statistical infrastructure the country has constructed for 200+ years made it possible to do that. Now we need to figure out how we’re going to envision what’s next and pay for it.

Nancy Potok: My bottom-line message is that the Census Bureau is one of the country’s great resources. It is critically important in understanding where we’ve been and where we’re going as a nation. As the world changes, the Census Bureau needs to be given the resources, legal authorities, and tools to continue to be a trusted resource for understanding our country, economy, and people. The time to do that is now.