A new study, using NORC at the University of Chicago’s nationally representative AmeriSpeak^® panel, reveals who frequent federal data users are and what they want from the federal data systems that inform policy, business, and research nationwide. We define frequent users as those who use federal data, including tables, reports, maps, summaries, and microdata, at least quarterly.

We found that most of them use accessible reports and tables rather than technical or restricted products, and that they trust federal statistics much more than the general public, though some concerns around privacy, confidentiality, and timeliness remain. Levels of trust vary by educational background and data use type: advanced degree holders and microdata users show the strongest confidence, suggesting that transparency about methods and safeguards may meaningfully boost users’ trust across the board.

Our survey, which expands on analysis from two recent American Statistical Association (ASA) reports, also found a surprisingly broad set of real‑world uses for federal statistics far beyond policy and research. These findings point to three investment opportunities—targeted measurement to capture evolving needs, clearer communication to reach different kinds of users, and engagement of applied users from a variety of backgrounds in modernization efforts—that could strengthen the federal statistical system’s resilience and public value.

Federal statistics underpin government decision-making, business planning, and research. From employment reports to census data and crop forecasts, federal statistics are used to allocate federal funds, apportion Congressional seats, guide policy decisions, and inform business planning and investment.

The two ASA reports (2025a, 2025b) document growing concerns about the resilience of the federal statistical system, citing sustained funding pressures, staffing constraints, and heightened political scrutiny. To inform investments in the system, we sought to understand how federal data users engage with the data, the value they derive from them, and how well federal data serve their diverse needs.

Expanding on the ASA reports, we used a unique nationally representative sample to understand Americans’ data use and views of federal statistics. Our findings center on frequent users: who they are, how they use federal statistics, where their confidence in federal data is strong, and where they have concerns. Public confidence hinges on whether the statistical system can remain accessible, timely, and credible, making these findings critical for future investment decisions.

Using NORC’s AmeriSpeak panel, a nationally representative, probability-based survey of U.S. adults, we collected data monthly from June 2025 through January 2026. AmeriSpeak allows consistent measurement over time, reliable comparisons across groups, and subgroup analysis, such as frequent data users or occupational categories. The panel also can be adapted for future targeted studies.

• 8,959 responses (approximately 1,110 per monthly wave):
  • 788 responses from 489 frequent data users
  • 8,171 responses from 4,766 other adults
• Nationally representative probability-based sample
• Multi-mode: web and telephone interviewing
• Tailored, expert-reviewed questions on federal statistics covering topics of trust, relevance, accuracy, timeliness, granularity, and privacy protection

Our study finds that frequent federal data users represent nearly 1 in every 10 U.S. adults. They have high educational attainment, with many working in applied, decision-oriented occupations such as education, science, business, and management.

Consistent with ASA findings on the diversity of federal data users, we found that most frequent users rely on reports and summaries rather than restricted or confidential microdata. Specifically:

• 70 percent cite official facts and figures from published reports 
• 61 percent use published tables and maps 
• 29 percent work directly with microdata 
• 14 percent use restricted or confidential data 

The takeaway: Sustained engagement with official statistics does not require secure access or advanced technical skills. Rather, well-documented, accessible statistics are essential to support data users’ real-world decision-making across occupations and business sectors. 

To capture the full spectrum of federal data uses, we included open-ended questions in the survey. Figure 1 shows a word cloud that draws from 437 responses collected across all survey waves. The size of the words reflects their relative frequency in respondents’ open-ended descriptions of their personal uses.

The most common theme is research—spanning academic research papers, graduate school projects, and college coursework—but respondents also describe highly practical applications: tracking crime rates in prospective neighborhoods, monitoring inflation and unemployment rates for personal investment, and consulting census data for business planning and marketing strategies. Respondents report many health-related uses, including public health programs, medical research, and COVID-related data queries.

According to our study, the estimated percentage of Americans who tend to trust federal statistics declined from a range of 55 to 57 percent in June and August 2025 to a range of 49 to 51 percent in December 2025 to February 2026. However, people who regularly use federal data report substantially higher levels of trust.

Figure 2 compares responses from frequent data users and the general public across a range of topics. It shows that an estimated 73 percent of frequent data users tend to trust federal statistics compared with 51 percent of the general public. Frequent data users are also more likely to agree that federal statistics are important for policymaking and business decisions.

Both groups report lower levels of agreement with statements about federal statistical agencies’ privacy protection and the timeliness of federal statistics. This suggests that focusing on transparency, clear communication about existing safeguards, and improvements in timeliness are system-wide priorities that can help the federal statistical system better serve data users. The gap between trust in statistics overall and confidence in privacy protections signals an opportunity for federal agencies to consider how they communicate the care they take to protect confidentiality. Federal data users’ responses about timeliness indicate that this will be an important area for future agency investments. 

We found that frequent data users with advanced degrees report markedly higher confidence in federal statistics than those without. There are striking differences regarding perspectives on confidentiality safeguards, perceptions of accuracy, and recognition of data’s value for decision-making—areas where clear communication and transparency could directly address users’ concerns.

Among frequent users with a graduate degree, 56 percent say they trust federal confidentiality protections, compared with 40 percent of those without graduate degrees, a gap mirrored in other areas. For instance, 75 percent of data users with graduate degrees agree that statistics are generally accurate, compared with only 54 percent of other frequent data users.

Of users with graduate degrees, 93 percent believe policymakers need federal data, compared with 67 percent of those without graduate degrees. These differences highlight a concrete opportunity for agencies to consider how they communicate about federal data quality to data users from different educational backgrounds. 

Table 1 shows that among frequent data users, those who engage directly with confidential or individual-level data—microdata and restricted-access records from surveys, censuses, or administrative datasets—show notably higher confidence in federal statistics than those relying exclusively on aggregated tables and reports. Microdata users are more likely to agree or strongly agree that federal statistics are accurate and have greater confidence in privacy protections.

While there are data users in both groups who would like more timely or granular data, microdata or restricted data users tend to have more positive perceptions of these aspects of federal data. For example, microdata users are more likely to agree or strongly agree that federal statistics are available with the subgroup detail they need (62 percent for microdata users, 44 percent for report or table-only users). Microdata users are also more likely to agree or strongly agree that federal statistics are available in a timely manner (54 percent for microdata users, 41 percent for report or table-only users).

It is possible that microdata and restricted data users view federal data more favorably because they have become more familiar through their research with agencies’ approaches to protecting data confidentiality and developing more granular, timely data. Still, transparency investments such as plain-language documentation of methods, visible quality checks, and demonstrated confidentiality protections may build confidence across broader audiences.

We find a substantial base of frequent, applied users who value and generally trust federal statistics but have persistent concerns about privacy, confidentiality, and timeliness of federal data. While there are several possible explanations for the patterns we found, the findings suggest that direct exposure to methods and safeguards may demonstrably strengthen confidence among data users.

Well-designed studies can deliver timely, policy-relevant insight, helping agencies monitor change, target interventions, and show the value of federal statistics even under constrained budgets. Our findings show that investment in understanding user needs can be an essential management tool for guiding decisions about the nation’s statistical infrastructure.

Our study points to opportunities for tailored measurement and analysis—tools that can help federal agencies, statistical system leaders, and stakeholders better serve diverse users and strengthen communication about the value and quality of federal statistics.

Data for this study were collected using NORC at the University of Chicago’s AmeriSpeak^® panel for a total of 8,959 responses between June 2025 and January 2026, with both web and telephone interviews. AmeriSpeak is a probability-based panel designed to represent the population of adults in the United States. We analyzed 317 additional open-ended responses from the February 2026 AmeriSpeak Omnibus. All interviews were conducted in English.

A technical overview of the AmeriSpeak panel is available on norc.org. Standard errors for analyses reported account for sampling error as well as the design effect due to survey weighting procedures. Survey results are subject to both sampling and non-sampling errors, such as nonresponse error and measurement error. There may be unmeasured error in this or any other survey.

Updated estimates, cycle-specific information on dates of data collection and sample sizes, and the questionnaires are available on the American Statistical Association website.