The Digital Divide Is Not Just About Access

When schools closed in March 2020 and instruction moved online, the conversation about the digital divide shifted in a way it had not in decades. Suddenly the question was not abstract. Students without reliable home internet or devices fell behind in ways that were measurable and documented. Districts scrambled to loan Chromebooks and work out hotspot arrangements. The immediate fix was access, because that was the crisis in front of them. But the students who had a phone and a hotspot still were not in the same position as the students who had a laptop, a quiet room, a parent who could help with the assignment, and a school-issued device with the right software. The access intervention was real and necessary. It was also not enough.

The digital divide has been studied and discussed in IS for decades, and for most of that time the framing was about access. Pew Research and others have consistently documented access disparities by income, age, rural and urban location, and race. I am not going to cite specific percentages here because those numbers move and my read of the Pew data may be out of date by the time anyone encounters this post. The directional pattern is not in dispute: lower-income households, older adults, people in rural areas, and some racial and ethnic groups have historically had lower rates of home internet access and device ownership. That gap has narrowed but not closed.

The access frame produced a particular policy response: subsidized broadband, school computer programs, library internet access. These are real interventions with real effects. But they rest on an assumption that the main thing preventing people from benefiting from the digital economy is the physical connection. My read is that this assumption misses most of the problem.

Having internet access and being able to use the internet effectively to improve your life are different things. Digital literacy, knowing how to search effectively, evaluate sources, navigate government and bureaucratic systems online, use productivity tools, manage privacy settings, identify scams, and communicate professionally in digital contexts, requires ongoing learning and practice. It is not automatically acquired once the connection is in place. And it is not uniformly distributed even among people with equivalent access.

The device quality dimension matters too. A phone is not equivalent to a laptop for most of the tasks that determine educational and economic outcomes. You can browse on a phone. Writing a research paper, building a spreadsheet, submitting a government benefits application, completing an online job application with document uploads, all of these are significantly harder on a small screen with a touchscreen keyboard. When we say someone has "access to the internet," a phone with a data plan counts. But the gap between that and a reliable home broadband connection and a dedicated laptop is enormous in terms of what is actually possible.

There is also what I would call a second-order divide, between people who can use technology and people who can build or modify it. Understanding how search algorithms rank results, how social media feeds are curated, how data is collected and used, and how to configure privacy settings requires knowledge that is not uniformly distributed and not naturally acquired through use alone. The person who uses a search engine daily may have no model of how results are ranked. The person who uses a social media app may have no model of what data the app collects or how that data shapes what they see. This is not a failure of intelligence. It is a failure of education and of system design. Systems are rarely built to be transparent about how they work.

This matters because people who lack that second-order literacy are not just less able to benefit from the digital economy. They are more exposed to its downsides. They are more likely to fall for misinformation designed to exploit engagement-optimizing feeds. They are more likely to be affected by algorithmic decisions they cannot interrogate, contest, or even identify. The digital divide, understood only as an access problem, misses the population of people who have access but are subject to the decisions of systems they do not understand and cannot influence.

IS researchers have been developing frameworks for thinking about digital skills as a spectrum rather than a binary. The bottom of the spectrum is basic operational competence: can you use a device and connect to the internet? Above that is informational literacy: can you find, evaluate, and use information effectively? Above that is communication and collaboration: can you participate in digital work contexts? Above that is something like critical digital literacy: do you understand how digital systems work well enough to make informed choices about how you use them? And above that, though most frameworks stop before this, is creation and modification: can you build, configure, or alter digital systems? Most policy discussions about the digital divide focus on the first level and occasionally the second. The third through fifth levels are where most of the economic and civic value of digital participation actually lives.

COVID made this visible in another dimension. Remote work shifted from an accommodation to a default for large portions of the workforce. People who could not manage their own technical setup, troubleshoot connection problems, navigate collaboration platforms, and communicate effectively in asynchronous digital contexts were at a significant disadvantage. That disadvantage fell along the same lines as the access gap but was not reducible to it. Two people with the same broadband connection had very different remote work experiences depending on their digital skill levels.

I think the access frame persists partly because access is easier to measure, easier to provide, and easier to claim credit for. A program that gives laptops to low-income students is legible in a way that a program to build sustained digital literacy is not. The second kind of program is slower, requires ongoing investment, and produces outcomes that are harder to attribute. That does not make the first kind of program wrong. It makes the first kind of program insufficient, and calling it a solution to the digital divide is a category error.

The divide is not between people who have internet and people who do not. It is between people who can fully participate in a digital society and people who cannot. The connection is a necessary condition for that participation, not a sufficient one.