Why We Need More Research on How Technology Affects Civic Engagement

The Research Gap That Shapes Public Power

When a county replaces paper voter registration cards with an online portal, when a state benefits agency moves applications to a mobile-first interface, when a city council adopts a digital comment platform for zoning hearings—each of these is a claim about civic engagement. The claim is usually implicit: this technology will make participation easier, faster, more inclusive. But the evidence base for those claims is thin, fragmented, and rarely tied to the procurement contracts that fund the systems. That is the research gap I want to name here.

Civic engagement, for the purposes of this field guide, means the ways people interact with public institutions: voting, commenting on rules, applying for benefits, serving on boards, contacting representatives, organizing neighbors. Technology affects each of these through specific mechanisms—identity verification, interface design, data matching, uptime, language access, and the procurement choices that determine which vendors build and maintain the systems. The gap is not that no one studies this. The gap is that the studies rarely connect to the contracts, audits, and oversight records that would let a practitioner act on them.

This matters because public digital systems are not neutral pipes. They are designed artifacts with political consequences. A 2023 report from the U.S. Government Accountability Office found that federal agencies often lack consistent measures for how digital services affect access to programs. A 2020 study in the Journal of Information Technology & Politics reviewed e-participation platforms and found that most evaluations focus on user satisfaction rather than whether participation changed policy outcomes or who was excluded. Those are two different questions. The second one is the one oversight bodies should be asking.

What We Actually Know—and What We Don’t

Identity verification as a gate

Identity verification is one of the clearest mechanisms by which technology shapes civic engagement. When a state unemployment agency adopts a new identity-proofing vendor, the design choice is not just about fraud prevention. It is about who can reach the benefit. In 2021, the Massachusetts Department of Unemployment Assistance faced a lawsuit over its identity verification system after claimants reported being unable to pass automated checks. The court record and subsequent agency data showed that certain claimants—particularly those without stable broadband, those with limited English proficiency, and those whose records had errors—were disproportionately affected. The mechanism is straightforward: a mismatch between a person’s data and a vendor’s reference data becomes a denial of access.

What we lack is a consistent, public dataset that tracks these denials across states and programs. The research question is not whether identity verification can reduce fraud. It is: what is the false-negative rate, and who bears the cost of those errors? Agencies rarely publish that number. Vendors treat it as proprietary. Auditors sometimes find it, but the findings are buried in PDFs that few practitioners read.

Algorithmic decision systems in benefits and criminal legal services

Algorithmic decision systems—rules-based or statistical tools that recommend eligibility, risk, or prioritization—are now embedded in public benefits administration and pretrial risk assessment. The research on their civic engagement effects is uneven. We have strong evidence that these systems can encode historical disparities. ProPublica’s 2016 analysis of the COMPAS risk assessment tool found that Black defendants were more likely to be flagged as high risk than white defendants with similar criminal histories. That finding was contested by the vendor and later reanalyzed, but the core question remains: how does an algorithmic recommendation change a human decision-maker’s behavior, and how does that change a person’s ability to participate in public life?

What we need is research that follows the decision chain: from data input to model output to caseworker override to client outcome. A 2022 audit by the New York City Department of Investigation examined the city’s benefits screening tool and found that caseworkers often did not understand how the tool generated its recommendations, leading to inconsistent overrides. That is a civic engagement problem. If a person cannot understand why they were denied, they cannot effectively appeal. The appeal process is a form of civic participation, and it is being mediated by a system that few people can explain.

Election administration resilience

Election administration is another domain where technology and civic engagement intersect. The shift to electronic poll books, ballot marking devices, and online voter registration has changed how people vote and how administrators verify eligibility. The research on these systems tends to focus on security and accuracy. Less common is research on how these systems affect voter confidence and turnout, especially among voters with disabilities, voters with limited English proficiency, and voters in precincts with older equipment.

A 2020 report from the Brennan Center for Justice documented that jurisdictions with newer voting equipment did not consistently see higher turnout, but the report noted that equipment age and poll worker training varied widely. The mechanism at work is not just the machine. It is the interaction between the machine, the poll worker, and the voter’s prior experience. We need more research that treats election administration as a civic engagement system, not just a technical one.

Why the Research Gap Persists

Three structural reasons explain why we lack the research we need.

First, procurement records are hard to link to outcomes. A city signs a contract with a vendor for a digital comment platform. The contract specifies features, uptime, and price. It rarely specifies what civic engagement outcome the platform is supposed to achieve, and it almost never includes a requirement to publish data on who used it and who did not. Without that link, researchers cannot connect the system to its effects.

Second, agencies have little incentive to measure exclusion. An agency that publishes a high false-negative rate for identity verification invites political scrutiny. An agency that publishes nothing faces no such risk. This is not a conspiracy; it is an incentive structure. Oversight bodies can change it by requiring access metrics in contracts and by funding independent evaluations.

Third, the research community is siloed. Computer scientists study system performance. Political scientists study participation. Public administration scholars study implementation. Each group has its own journals and conferences. The questions that matter—how does a specific design choice change who participates—fall between the silos. Interdisciplinary work is possible, but it is not rewarded in the same way as single-discipline work.

What Better Research Would Look Like

I am not calling for more studies that survey users about satisfaction. I am calling for research that meets four criteria.

1. It is tied to a named system and a named contract.

Instead of “e-government and trust,” study the specific portal, the specific vendor, the specific contract terms, and the specific population. Name the system. Name the agency. Name the oversight body that has jurisdiction. This is how you build a body of evidence that practitioners can use.

2. It measures exclusion, not just inclusion.

Most evaluations report how many people used a system. Fewer report how many tried and failed, how many were deterred by identity verification, how many could not access the interface in their language, how many gave up after a timeout. Those are the numbers that matter for civic engagement. They are also the numbers that agencies are least likely to publish voluntarily.

3. It uses administrative data and FOIA records.

Surveys have their place, but they are expensive and often unrepresentative. Administrative data—application logs, denial reasons, call center records—can show patterns at scale. FOIA requests can surface vendor performance reports and internal audits. Researchers should treat these records as primary sources, not as background.

4. It includes the people affected.

Research on civic engagement should not be done to communities. It should be done with them. Community organizations, legal aid groups, and disability rights advocates often have the clearest view of how a system fails. Their knowledge should shape the research questions and the interpretation of findings.

What Practitioners Can Do Now

You do not need to wait for a comprehensive research agenda. If you sit on a procurement committee, an oversight board, or a community advisory group, you can raise the questions that force better data into the open.

Ask: What is the false-negative rate for identity verification, and how is it measured? Ask: What data will the vendor provide on who uses the system and who does not? Ask: How will we know if this system excludes people, and who will be responsible for fixing it? Ask: Will the contract require public reporting of access metrics?

These questions are not technical. They are governance questions. They belong in the procurement record, the oversight hearing, and the public comment period. The research gap is real, but it is not an excuse for inaction. Every contract is an opportunity to generate the evidence we need.

FAQ

What is the main research gap in technology and civic engagement?

The main gap is the lack of studies that link specific public digital systems—and their procurement contracts—to measurable civic engagement outcomes, especially exclusion. We have many studies on user satisfaction and few on who fails to participate and why.

Why does identity verification matter for civic engagement?

Identity verification is a gate. If a person cannot pass an automated check, they cannot vote, apply for benefits, or comment on a rule. The design of that gate—what data it uses, how it handles errors, how it communicates failures—determines who is included and who is shut out.

How can oversight bodies improve research on this topic?

Oversight bodies can require access metrics in contracts, fund independent evaluations, and publish the results. They can also use their subpoena power to obtain vendor performance data that agencies do not release voluntarily.

What is one question I can ask in my next procurement meeting?

Ask: “What is the false-negative rate for this system’s identity verification or eligibility screening, and how will we measure it before and after deployment?” That single question can change what gets written into the contract.

Related Reading

This article is part of a recurring field guide on procurement accountability for public digital systems. Future pieces will examine specific contract clauses that govern access metrics, the role of community advisory boards in system design, and how to read an agency’s FOIA response for evidence of exclusion. If you work in oversight, procurement, or community advocacy, the next step is to request the access data for a system you already use. The research gap is not just an academic problem. It is a governance problem, and it is one you can help close.