Broadband internet access has moved, within roughly a decade, from a convenience to something researchers and policymakers increasingly treat as essential infrastructure on par with electricity or water service. That shift accelerated sharply during the pandemic, when broadband access determined whether children could attend school, whether adults could work remotely, and whether patients could access telehealth, exposing just how many American households, disproportionately rural, tribal, and lower-income, lacked reliable connectivity. The policy response has been substantial: tens of billions of federal dollars are now flowing toward broadband expansion, representing one of the largest infrastructure investments of its kind in American history. Whether that investment succeeds in closing the digital divide depends on lessons from research that predates the current funding wave.
Three Distinct Problems, Not One
The digital divide, as researchers have come to understand it, is actually at least three distinct problems layered together, and policy that addresses only one of them tends to underperform. The first is physical infrastructure availability: many rural and remote areas simply lack broadband infrastructure because the cost of running fiber or cable to sparsely populated areas exceeds what private providers can profitably recoup, a classic market failure that justifies public subsidy. The second is affordability: broadband infrastructure may be physically available, but the monthly cost puts it out of reach for many lower-income households, a problem concentrated in urban areas that otherwise have ample infrastructure. The third is adoption, sometimes called the digital skills gap: even where broadband is both available and affordable, some households, often older adults or those with limited formal education, do not subscribe or use available service effectively due to unfamiliarity with the technology or lack of a compelling perceived need.
Much of the historical federal investment in broadband, going back to programs beginning in the early 2000s, focused almost exclusively on the infrastructure availability problem, funding grants and loans to expand physical broadband networks into unserved areas. Research evaluating these earlier programs has found genuine but limited success: infrastructure expansion programs did measurably increase broadband availability in targeted areas, but subscription rates in newly served areas often lagged well behind availability rates, suggesting that solving the infrastructure problem alone left substantial gaps related to affordability and adoption unaddressed. This finding has directly shaped design of the current generation of federal broadband programs, which increasingly pair infrastructure funding with affordability subsidies and digital literacy initiatives, reflecting the research consensus that a multi-pronged approach outperforms infrastructure investment alone.
Why Broadband Maps Got It Wrong
Mapping accuracy has emerged as a surprisingly consequential technical policy issue. Earlier federal broadband maps, which determined which areas qualified for infrastructure funding, relied on data reported by internet service providers themselves and used a methodology that counted an entire census block as served if even a single household within it had access, a standard researchers and local officials widely criticized as dramatically overstating actual coverage, particularly in rural areas where a single served address could mask dozens of unserved neighbors spread across a large geographic block. Revised mapping methodologies adopted more recently use more granular, location-specific data, and early analysis suggests this has meaningfully improved the accuracy of funding allocation, directing resources toward areas that earlier maps had incorrectly classified as already served.
The affordability piece of the puzzle has its own research base. A federal subsidy program providing discounted broadband service to low-income households demonstrated, while it was funded, measurable increases in broadband subscription among eligible households, with researchers finding particularly strong uptake among households that had previously cited cost as the primary barrier to subscribing. When funding for that program lapsed, researchers tracking affected households found meaningful numbers either downgraded their service plans or dropped internet service entirely, offering a fairly direct piece of evidence that affordability subsidies, not just infrastructure investment, produce measurable changes in connectivity for low-income populations.
The Thin Evidence on Digital Literacy
Digital literacy and adoption programs, the third leg of the digital divide, have the thinnest research base of the three, partly because they are harder to evaluate at scale and partly because they have historically received the smallest share of federal broadband funding. The evidence that does exist suggests targeted digital skills training, particularly programs embedded in trusted community institutions like libraries, senior centers, and community colleges, can meaningfully increase both broadband adoption and effective use among populations who have infrastructure access but have not historically subscribed or used it confidently.
Taken together, the research suggests that closing the digital divide durably will require sustained attention to all three dimensions of the problem simultaneously, rather than treating infrastructure buildout as a one-time project that, once complete, resolves the broader access gap. Given the scale of current federal investment, researchers studying broadband policy will have an unusually rich dataset over the coming years to assess which combination of infrastructure, affordability, and adoption strategies actually delivers durable, equitable connectivity, and which elements of the current approach may need to be redesigned as implementation proceeds.
