To Improve Broadband Deployment, Enhanced Data Collection Is Key

Overview

The federal government has a long track record of significant investment in expanding broadband to ensure that all Americans can access the internet. A 2022 Government Accountability Office (GAO) report detailed some 133 programs across 15 federal agencies that have supported broadband deployment efforts. These programs, which provided funding to construct new networks and programs focused on household affordability and digital skills training, cost $44 billion from 2015 to 2020 alone. After that period, the federal government made further investments as part of its COVID-19 pandemic recovery response through programs—such as the Treasury Department’s Capital Projects Fund—that helped people access critical health services, education, and employment online.

In addition, the Infrastructure Investment and Jobs Act (IIJA) of 2021 is expected to deliver over $65 billion for broadband programs by 2030, primarily through IIJA’s Broadband Equity, Access, and Deployment program (BEAD), which is designed to bolster broadband deployment and help more people use the internet to improve their lives. (Note: The Digital Equity Act, known as DEA, which also was intended to boost programs tied to adoption and digital skills gaps, was canceled by the Trump administration in May 2025.)

These investments show that the federal government is focused on expanding broadband access by funding not only network deployment but also programs targeting issues, ranging from affordability to digital skills, that are barriers to consumers beginning to use broadband. This work is ongoing, as bipartisan working groups in Congress are considering reforms to the Universal Service Fund (USF)—a fund accounting for roughly $8 billion in annual federal broadband expenditures. Such reforms could alter the federal funding available for broadband deployment.

Yet broadband policy remains poorly equipped to target funding, identify existing gaps, and evaluate different approaches to achieving universal access despite federal investments in data collection. These include a mandate from Congress for the Federal Communications Commission (FCC) to “collect standardized, granular data on the availability and quality” of both fixed and mobile internet services and make this information publicly available. There is also an effort by the FCC and the National Telecommunications and Information Administration (NTIA) to develop standardized broadband mapping data nationwide. These limitations in federal data collection make it more difficult for policymakers, researchers, and local-level broadband advocates and nonprofits to consistently identify and address gaps in broadband availability, adoption, and affordability.

Select demographic data about broadband users is available through the U.S. Census Bureau, and information about who is using broadband—including basic questions about broadband subscriptions and home technology use—is provided by government agencies including the FCC, NTIA, and the U.S.

Department of Agriculture (USDA). But a pressing need remains for national datasets about broadband pricing, network quality, consumers’ digital skills, and more, regardless of which government agency might provide the information.

Incomplete information can make it difficult for lawmakers to understand the effects of policy, and for government entities to enforce grant recipients’ program requirements. Because federal data focuses on information like the number of households with subscriptions, it lacks feedback on network quality and reliability, or on how consumers experience a broadband connection. Incomplete data can also make it challenging for researchers to evaluate how specific broadband policy changes affect consumers, and for people involved in broadband deployment and adoption at the state level to monitor whether ongoing and new programs are meeting their goals. Lack of data on broadband’s societal impact—such as effects on the economy, education, and health care—has led states to conduct their own baseline measurements to meet BEAD program requirements, including BEAD’s aim to “grow economic opportunities … provide increases in access to health care services … and enrich education experiences” nationwide. Additionally, NTIA’s BEAD Five-Year Action Plan guidance requires states to evaluate how program implementation affects economic and workforce development, education, health, civic and social engagement, and access to essential services.

To highlight shortcomings in the federal data, The Pew Charitable Trusts assessed the literature on broadband that relies on federal datasets, in some cases going back decades. The assessment is not a comprehensive review of broadband data and research; instead, it is designed to highlight the most consequential data gaps as identified by leaders in the field.

Data collection issues have taken on new urgency because the BEAD program places significant reporting requirements on state broadband offices that must be met by 2030 and need accurate data to evaluate the successes and challenges of their efforts. On its own, BEAD requires states to collect data about broadband access and adoption, and to demonstrate the impact of both on measuring outcomes related to economic growth, education, health care, and civic participation. Yet federal agencies have not provided best practices for collecting this data or for engaging with the public on it. State approaches to BEAD implementation may differ, but clear data and measurement standards from the federal government would enable consistency and comparability across states.

This brief not only outlines Pew’s review of the broadband literature assessing the gaps in available data but also looks at the available federal data sources for broadband. Finally, it explores the public policy implications of these gaps—and suggests some possible solutions to address them.

Pew’s review of broadband literature published from 2008 to 2024 (based on data that, in some cases, dates to the 1990s) showed that researchers had identified the following shortcomings with broadband data:

• Limitations in geographic and household data: Federal data on broadband access, adoption, and household characteristics is often reported by county, ZIP code, or census tract, rather than by household. And geographic and household characteristics—such as the regions used in data collection or the composition of households—are not standardized across data sources, making it difficult to combine datasets for evaluation. This lack of standardization has led to inaccuracies in measuring the impact of broadband and the gaps in coverage, which is particularly problematic for federally funded projects that rely on accurate mapping data for broadband investment; these projects could misappropriate funds because of the data’s shortcomings.

   

• Reliability of federal data for evaluation purposes: Federal data relies heavily on information from internet service providers (ISPs), leading to concerns in the literature that data on broadband availability and network performance may be affected by bias and lack of transparency. Also, federal data tends not to include price, making it difficult to assess how much broadband service costs and how affordable it is for customers throughout the country.
• Inconsistencies in definitions: Several federal sources do not specify which broadband technologies are available to consumers, making it difficult to correlate network performance and household use of the internet. Some sources have also been inconsistent in defining key components, ranging from what qualifies as a broadband connection to the meaning of the phrase “digital equity” to how to measure internet use and digital skills—and how consumers benefit from the connection.
• Lack of consistent impact assessments: The available data makes it difficult to determine the availability of affordable broadband connections or how to define affordability, and there is no uniform method of tracking broadband adoption patterns over time. Researchers need more information to gauge the value and impact in multiple areas (economic development, health care, education, civic life) of having a broadband internet connection.

What national data sources exist?

Broadband availability data

Since 1992, the federal government has collected data on broadband availability, affordability, and adoption. The data about broadband availability has focused on the number of locations, such as homes and businesses where broadband service is available; the number connected to the internet; the technology type used; and the companies that provide those connections. From 1999 to 2022, availability data was collected using the FCC’s Form 477, which included information on “broadband connection; wired and wireless local telephone services; and interconnected Voice over Internet Protocol (VoIP) services in the 50 states, the District of Columbia, and U.S. territories.” Starting in 2022, this information was collected by the FCC’s Broadband Data Collection System. Broadband availability data is also provided through the National Broadband Map project, which was managed by NTIA from 2010 to 2014 and then by the FCC. This map displays where services are provided, using information that ISPs report to the FCC.

One of the main limitations in federal data about where broadband is available is a lack of precision about the number of locations served within a geographical area, whether at the census block (the smallest geographic area in which the U.S. Census Bureau collects data), ZIP code, or county level. Other limitations are detailed below.

Broadband pricing and affordability data

Even when broadband service is available, consumers are often unable to afford it. Yet data about affordability is limited. Some information is available through the FCC’s Urban Rate Survey, which is an annual sample of ISPs operating in urban census tracts and focuses on the prices charged by ISPs across speed tiers and technology types. While this survey is a valuable tool for monitoring price changes over time, it by definition excludes rural areas and does not provide researchers and policymakers with the final price that consumers ultimately pay for internet service, making affordability difficult to assess.

Similarly, the Bureau of Labor Statistics’ Consumer Price Index details urban consumers’ average internet service costs. However, like the Urban Rate Survey, this data reflects only urban consumer patterns and prices and doesn’t apply to rural broadband pricing, which historically has been higher than in urban areas. Additionally, rural residents, on average, have lower incomes than their urban counterparts, meaning a reliance on the Consumer Price Index, Urban Rate Survey, or any other urban-centric broadband pricing source risks underestimating the affordability challenges facing rural Americans and distorting any state or national analysis of price and affordability.

But there’s some good news on pricing: As of October 2024, all internet providers must participate in the FCC’s Broadband Consumer Labels program. These labels, much like nutrition labels for food, display— online or in stores where consumers would order internet service—base prices, speeds, fees, data allowances, and other plan information. Although it is a welcome step toward transparency, the program is so new that it’s not yet possible for researchers to evaluate its effectiveness—or to analyze the impact of the labels on affordability and pricing. So far, no plans have been announced to aggregate pricing data obtained through the Broadband Consumer Labels program.

Broadband adoption and usage data

Once broadband service is available in an area, it is important for policymakers and researchers to understand how consumers take advantage of access to connections. As detailed by NTIA’s Internet Use Survey, there are several reasons individuals might remain unconnected, including price and a perceived lack of need. Data collection on broadband adoption can help illuminate those barriers and which ones, beyond access to a connection, might be preventing consumers from getting online. The FCC’s Form 477 data includes broadband adoption data per 1,000 households by census tract going back to 2000.

Demographic data, used to analyze trends in broadband adoption by age, race, income, and more, is available from two separate U.S. Census Bureau surveys: the American Community Survey and the Current Population Survey. Beyond household composition, the American Community Survey includes estimates of the percentage of homes with a broadband connection down to the level of a census tract, each of which typically includes roughly 4,000 people. NTIA’s Internet Use Survey also maintains survey data regarding internet use and adoption patterns at the national and state level.

However, as with broadband availability data, adoption and usage data is limited by the geographic sublevels (state, county, ZIP code, census tract) of information about household and business connectivity patterns, and available datasets can vary at geographic levels, so they don’t always align for analysis. This means that the adoption and usage data is not helpful in evaluating household-level decisions about broadband connectivity, which could lead to flawed assumptions in policies designed to address disparities in connectivity.

Federal program funding data

Given the significant, ongoing federal investment in broadband expansion, assessing program effectiveness and promoting transparency rely on an understanding of how federal funds are distributed and spent. NTIA maintains several lists detailing the federal funding available to increase broadband deployment, adoption, and affordability—including the Public Geographic Information System (GIS) Data site, which details the funding type and award amount for programs the agency administers. The agency also maintains a Federal Broadband Funding Dashboard, which displays all federal agency and program funding for broadband from 2020 onward. However, NTIA has acknowledged inaccuracies in the dashboard about broadband-specific funding across agencies, as well as an inability to connect federal funding awards to how states choose to use the funds.

The Universal Service Administrative Co. (USAC), an independent, not-for-profit company that administers the Universal Service Fund for the FCC, provides historical data about two subsidy programs—Lifeline, a $9.25 monthly federal subsidy to low-income consumers (those with income less than 135% of the federal poverty guidelines), and the now-defunct federal Affordable Connectivity Program, which until its cancellation in June 2024 provided up to $30 per month to qualifying households for an internet subscription. This data includes participation rates, total funding amounts, and qualification rates for both programs, but only down to the county level, so its use in evaluation is limited by the lack of precision about how receiving broadband subsidies affected households.

Additional datasets

Pew also examined broadband adoption data alongside various federal data to assess the impact of broadband deployment and connectivity—including economic data from the Bureau of Economic Analysis, the USDA and its Economic Research Service, and the U.S. Census Bureau. Table 1 summarizes these federal data sources.

The limitations of federal broadband data

Issue No. 1: Limitations in geographic and household data

One of the primary limitations that Pew identified in the literature was that historical data on where broadband was available was aggregated and reported at the census block or census tract level.¹ Measuring broadband adoption or access within large geographical areas has meant that some locations reported as served by broadband providers include households or businesses for which broadband service may not be available.

Table 2 illustrates inaccuracies in measuring broadband gaps, especially in rural areas. Initial mapping efforts by the FCC have highlighted discrepancies, often overlooking locations and showing a mismatch between the reported availability of broadband services and the availability experienced by users.²

Federal agencies have also shown a lack of standardized geographic definitions, with the U.S. Department of Agriculture offering 30 different definitions of the word “rural.” Such inaccuracies risk misdirecting broadband investments to regions not in need, given that policymakers need more clarity on where funding would be best directed.

There is also a shortage of household-level data about broadband access and adoption. This has led to uncertainty regarding actual coverage across the country.³ The National Broadband Map, which displays providers and technology types available at the household level, has begun to address this information gap, but more time is needed to fully evaluate its effectiveness.

Available household characteristic data has proven problematic on several fronts. For example, it is not possible to accurately account for network quality and usage data at the household level or by demographic groups.⁴ While the American Community Survey allows for a baseline measure of home broadband connections, critical components—such as the type of connection and household composition—still need to be included in the survey to help evaluate the impact of broadband subsidies, such as Lifeline and the now-defunct Affordable Connectivity Program.⁵

Issue No. 2: Reliability of federal data for evaluation purposes

A consistent theme in the literature was that federal data sources relied on self-reported information from internet service providers. Several studies found that ISP-provided data appears to overstate the availability of broadband services and advertised speeds compared with those identified by the research community, as determined by speed tests to determine the level of service that locations actually receive.⁶ Pew also found a need for more detail on the type of internet connection (fiber, cable, fixed wireless, DSL, or satellite) available to consumers.⁷ Without reliable data on coverage and speed, it can be difficult for key decision-makers to identify the extent of needs across the country, or to understand how best to target resources to close gaps. The scope of this challenge was recognized by the Congressional Research Service (CRS) in 2022, when it warned that “inaccurate broadband deployment data … could affect the share of BEAD Program funding an eligible state or territory receives.” While this CRS report focused on BEAD, similar issues could plague future federal and state funding.

Pew researchers also noted delays between collecting federal data and reporting it, suggesting that information about broadband deployment and broadband use, which is currently reported once a year, may need to be updated more often.⁸ The timing question is particularly problematic for any analysis that attempts to assess impact over time, as there is sometimes a significant delay between when broadband investments are made and when their effect on adoption and other socioeconomic indicators can be observed.⁹

Pew also identified the lack of accurate pricing data as a significant barrier for evaluating the impact of broadband affordability programs.¹⁰ Even when pricing data is available through the FCC’s Broadband Labels program, it may be difficult to estimate prices for individual households (as opposed to regional averages); to assess whether a price applies only to bundled packages (internet, television, and phone services); or to accurately reflect network quality.¹¹

Issue No. 3: Inconsistencies in definitions and consumer input

One significant issue with federal data has been the shifting definitions across agencies of broadband-related concepts, such as speed thresholds, which also include a lack of consistency in distinguishing between technological differences in broadband connectivity.¹² Although providing a baseline assessment of connectivity that does not distinguish between the technology used for a connection is valuable, there are strengths and limitations associated with each technology that remain unaccounted for in any evaluation or analysis relying on available federal data. Research shows a link between broadband adoption and economic growth, but its impact is hard to gauge without knowing the type of connection used.¹³ For example, evidence suggests that faster broadband speeds are associated with higher rates of economic growth among Organization for Economic Co-operation and Development (OECD) countries. However, the lack of more precise federal data from the American Community Survey, which does not distinguish among the connection speeds of whichever plans households have chosen, limits any national, state, or local analysis that could replicate the OECD data.

In addition, federal data does not include historical standards or definitions for digital equity or impact assessments, including assessments of broadband affordability and digital skills to monitor adoption and track trends in internet use. This limitation has affected BEAD implementation, forcing states to collect digital equity data, conducted well before the program’s cancellation in May 2025, without federal baselines or national benchmarks—making consistent nationwide assessment in areas like economic development, health care, education, and civic participation difficult.¹⁴ This inconsistency will also make advocating for policy changes difficult, given that the baseline values and measures of progress may not translate across the states.

Federal data often does not include consumer surveys about broadband connectivity, presenting a missed opportunity to understand local contexts and issues that can be overlooked in quantitative studies alone. While NTIA’s Internet Use Survey does provide some information, this data is limited to national and state-level aggregations and does not address households at a local level. Researchers have also made clear that although the Broadband Label program will provide more detail on service options, it will not provide insight into how consumers choose a broadband plan.

Issue No. 4: Lack of consistent impact assessments

The broadband literature has, over the years, identified several themes about the impact of having a broadband connection. One is that no federal data consistently tracks pricing and affordability across the country.¹⁵ The Urban Rate Survey and Current Population Survey detail average prices, speed tiers, and technology types for urban consumers but don’t contain such information about rural areas or include variations in affordability across the country. This is particularly problematic in the case of low-income areas that may have access to an internet connection but where some households may remain unconnected because of the cost. Additionally, this self-reported data from providers doesn’t reflect what individual households currently pay for broadband service.

The gap in pricing and affordability data feeds into the problem of federal data being unable to assess adoption patterns nationwide. Many factors affect adoption rates, including affordability, digital skills, and demographic and economic circumstances, but there are no federal datasets assessing adoption patterns that holistically account for multiple factors beyond broadband availability. This problem becomes more severe when researching minority communities. For example, researchers have done minimal work on Tribal broadband connectivity, and government data on both broadband infrastructure and adoption patterns in these communities remains limited.¹⁶

What’s more, no federal datasets attempt to directly address the ultimate value that a broadband connection brings to a household, including economic, health care, education, and quality-of-life improvements.¹⁷ The nature of the data collected—primarily quantitative and lacking qualitative evaluation methods—reflects this approach. This makes it difficult to establish causal relationships between connectivity and employment, for example, and broader economic value, particularly about economic sectors and business adoption patterns.¹⁸ Similarly, there has been limited assessment of broadband’s connection to education and public health outcomes, with minimal federal guidance on best practices to doing so.¹⁹

Implications of data gaps on broadband policy

Outlining the implications of these shortcomings for policymakers working on broadband expansion is critical to translating research findings into practice. Pew researchers have identified three broad shortcomings: the potential misallocation of public resources; limitations in connecting public investments to desired societal outcomes; and inefficiencies in program implementation and evaluation.

Potential misallocation of public resources and avoiding funds duplication

The current reporting structure and data make it difficult to avoid misallocating funds across the states. Previous reporting metrics, which relied on large geographic areas, often overstated the number of households and businesses with broadband access—leaving them ineligible for future broadband funds, because federal and state broadband programs are required to avoid funding the same area more than once.

For example, NTIA has directed that state BEAD program funding treat a location as “served” if it has already received federal or state broadband deployment funds, even if the area remains unconnected. Congress’ intent with IIJA was to avoid funding duplication, but the available data places a significant burden, both in time and financial resources, on states, consumers, and industry to determine which areas are eligible for funding. This burden also hinders the creation of targeted programs and policies that are needed to ensure that the regions most in need are the ones that receive funding.

Data quality limits analyzing the connection between public funding and desired societal outcomes

The unreliability of federal data also has significant implications for policy development and complicates efforts to connect performance obligation data with broader societal outcomes. State broadband offices need reliable data to monitor and evaluate broadband grant programs effectively and ensure that grant recipients meet their obligations for network performance and speed. Also, a delay in federal data creates obstacles for new programs and funding decisions that depend on information about availability and adoption rates. For example, because census data is updated on an annual basis, evaluation efforts and funding decisions could be made with outdated information.

Inconsistencies in definitions continue to plague program implementation and evaluation

The lack of alignment between federal broadband programs has also been exacerbated by the use of inconsistent definitions across the agencies and programs involved. As the 2022 GAO report details, the absence of a national broadband strategy has led to a fragmentation of federal broadband programs— hindering efforts to prevent waste, fraud, and abuse, and limiting the federal government’s ability to reduce program overlap and duplication of effort. At the very least, inconsistences in definitions and conflicting federal programs indicate that efficient program evaluation in broadband policy will remain a challenge until such discrepancies are resolved.

Such inconsistent definitions and program misalignment also make it difficult to measure outcomes related to broadband. With both BEAD and DEA, Congress took steps to address the problem of available federal data not being conducive to connecting broadband access and adoption to outcomes that are of interest to policymakers, including economic, health care, education, and quality of life issues. Implementation guidance from NTIA has also tasked states with measuring these components of broadband deployment. In fact, much of the burden for data collection has fallen to the states, with minimal guidance from the federal government on best practices or data standardization. The risk for policymakers is that future evaluation of outcomes tied to broadband deployment will run into the same issues of alignment and fragmentation as those identified by the GAO, such as the fragmentation of federal programs, and varying levels of data quality across states could hinder future evaluation efforts.

Pew’s assessment of policy changes to address the limitations in federal data

An example of the difficulties in assessing what policymakers want to know about broadband policy— and what’s knowable with current data—comes from NTIA’s Broadband Technologies Opportunities Program (BTOP) from 2009. Recognizing that affordability and digital skills are significant barriers to broadband adoption, BTOP provided discounts on computer equipment and broadband subscriptions, along with outreach efforts and digital skills training. However, as noted by the GAO, NTIA at the time failed to establish goals for BTOP and was unable to “assess the effectiveness of these approaches in addressing adoption barriers.” The inability to accurately evaluate the program sparked skepticism among lawmakers, with some subsequently regarding programs that provide devices, such as computers, phones, or tablets, as government waste and tax-funded giveaways. Although the BTOP program wasn’t fully assessed, later research highlighted the complex factors needed to make publicly funded broadband networks successful and boost economic benefits.²⁰ The BTOP program is just one example of the difficulties in assessing the effectiveness of public funds in broadband policy.

Other shortcomings became clear during the early implementation of BEAD, particularly regarding funding allocations and the mandate to avoiding funding duplication—a difficult task, given the scope of BEAD’s coverage and the fragmentation of federal programs. Tracking which areas received federal broadband funds has slowed BEAD implementation and highlights concerns about government waste, as poor data systems have hindered past projects.

If policymakers want to know the effect of broadband policy, they may need to pass legislation addressing affordability, network quality, and methods for measuring impact.

How policymakers can address affordability and competition policy

The glaring gap in broadband data about the quality and reliability of available pricing information limits the evaluation of broadband affordability and the impact of marketplace competition on prices nationwide. It also works against lawmakers’ intentions to enhance affordability and competition, particularly regarding new technologies.

And inconsistencies and potential biases result from the fact that, as mentioned repeatedly in the literature, federal sources rely heavily on pricing information reported by broadband providers. While the FCC’s Broadband Labels program, designed to provide transparency about service plans, is a step forward, its reliance on voluntary reporting and lack of aggregation limits its usefulness as an evaluation tool. The current pricing data doesn’t allow researchers to fully assess trends in broadband prices related to marketplace competition or to determine whether available plans are affordable to most American households, particularly those with low incomes. To effectively evaluate broadband affordability and competition, policymakers must invest in independent, standardized data collection that goes beyond provider-reported pricing.

Investments needed to improve data on network performance and quality

Evaluating broadband’s impact on outcomes tied to telemedicine, remote work, and education is challenging, as outcomes depend on network quality—yet federal data, like the FCC’s Broadband Map, mainly tracks availability, speed tiers, and technology types. While these indicators are essential, they don’t provide a holistic view of network performance or quality. For instance, advertised speeds often differ significantly from actual speeds, particularly during peak usage hours.

To answer questions about network performance, policymakers would need to devise reporting requirements that assess actual network quality, covering parameters such as average download/upload speeds, latency (the amount of time it takes data to travel to its destination and back along the network), and reliability. Federal policy could be modeled on existing state models—such as requiring ISPs receiving state funds to report data on network operations—and on take rates (the percentage of customers with access to a network who choose to subscribe). Consumer-reported data through crowdsourcing platforms could also be helpful, including initiatives like Speedtest by Ookla or Measurement Lab (M-Lab). Multiple solutions exist, and it would be up to policymakers to determine the best path forward in collecting network performance data to accurately assess the impact of broadband connectivity throughout the economy.

Funding needed to link broadband connectivity and outcomes

While BEAD requires states to establish benchmarks for broadband’s socioeconomic impact, the absence of federal data complicates these efforts. Particularly lacking are datasets focused on minority populations, including broadband issues in Tribal nations and populations targeted in the Infrastructure Investment and Jobs Act of 2021. Consequently, states have relied on fragmented or localized studies to fill in the data gaps, which may lead to uneven evaluation efforts regarding the program’s impact.

So, if policymakers, particularly those at the federal level, want to answer questions about the impact of broadband, they would need to direct federal or state funding to collect the necessary data to make those impact assessments possible. Collaborative efforts among agencies such as the FCC, the Census Bureau, and the Department of Labor could yield a more integrated dataset, similar to the ongoing collaboration between the Census Bureau and NTIA on the Local Estimates of Internet Adoption initiative, which is developing internet adoption estimates for smaller geographic areas overlooked in current federal datasets. Partnerships with academic institutions and research organizations—such as the continuing collaboration between The Pew Charitable Trusts and Michigan State University to create an evaluative framework for the broadband components of IIJA, and analysis of the current barriers to broadband connectivity by Purdue University’s Digital Divide Index—could also refine impact assessment methodologies.

An understanding of broadband’s nuanced effects on households and communities requires both quantitative data and qualitative research. Surveys capturing user experiences, digital literacy levels, and the perceived value of connectivity could provide policymakers with a richer context for decision-making, as noted by multiple scholars.²¹ And qualitative research not only adds evidence of how people are affected by policies and programs but also can be used, as articulated by researchers at the Brookings Institution, to “understand … why a given program or intervention may or may not work as intended”—in a way that cannot fully be captured in quantitative or survey data. Ultimately, providing funding for robust impact assessments, using both qualitative and quantitative data, would help validate federal investments and guide future initiatives to maximize broadband’s societal benefits.

Conclusion

Researchers and policymakers have made good use of federal data on broadband deployment and access. Evaluations of the data over the years have provided insight into program design and its effect on the public. However, shortcomings in federal data suggest that more could be done at the federal level to equip implementation and evaluation efforts. By providing the tools needed to evaluate the relationship between connectivity and socioeconomic outcomes, policymakers could build broader consensus across government and industry on the importance of sustained investments in broadband.

Research methodology

This project incorporates aspects of design practices from Petticrew and Roberts (2006), Foster and Jewell (2017), and Richter et al. (2020), among others. The project also relied on Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) materials, including research checklists and additional resources for conducting systematic reviews.

Generally, the analysis followed a seven-step process for evaluation:

1. Creating the conceptual framework: Researchers from Pew’s broadband access initiative (BAI) outlined the conceptual framework by focusing on two critical components: available federal datasets related to broadband and the limitations for research and evaluation as identified by the broadband research field. Framing these questions also led BAI to consider the consequences of data limitations beyond flawed research findings in the literature, including misaligned program goals and improper evaluation of current federal broadband programs.
2. Identifying the selection criteria: Next, BAI searched for academic literature from 2008 to 2024.This was done to encompass any results from the American Recovery and Reinvestment Act of 2009, as several federal programs from that era centered on broadband deployment. BAI researchers also wanted to include any research conducted during and after the COVID-19 pandemic, as significant federal investments were made during that era as well.
   
   This project intentionally includes only references from peer-reviewed academic journals relying on federal data sources that specifically list shortcomings in federal data identified by the researchers.
3. Developing the search strategy: Next, BAI worked directly with The Pew Charitable Trusts’ library and archives team to refine the search parameters. Based on an initial review, the search queries were made using the following keywords:
   • Broadband policy o FCC 477
   • National Broadband Map o American Community Survey o Broadband access
   • Broadband affordability
4. Selecting relevant studies: The literature review resulted in a list of 12,000 articles from 2008 to 2024 that relied on federal data. Initial searches were coordinated with Pew’s library and archives team, relying primarily on searches through EBSCO Information Services.
   
   BAI reviewed the articles on the initial list to select studies for this analysis. All reports were required to use federal broadband data in their analysis; be explicitly focused on U.S. broadband policy; be published in peer-reviewed journals; and discuss the limitations of federal datasets used in the study.
   
   Pew’s library and archives team narrowed the initial list to 500 articles. After reviewing for duplications and search relevance, the team selected 50 articles for the final analysis.
5. Coding the identified studies: Once the final selection was made, BAI conducted a coding process to assess which articles addressed the primary research questions for this study. Four general themes emerged through this process.
6. Critically appraising the identified studies: Following this review, BAI conducted an internal review of the identified studies to further assess their appropriateness for use in the analysis. This was done in conjunction with peer reviewers outside The Pew Charitable Trusts.
7. Synthesizing the data: Finally, BAI synthesized the data based on the themes identified. Our general conclusions are based on the preceding analysis, and the proposed policy changes are based on the shortcomings observed in federal data.