IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Including People With Disabilities in Data Collection

Government and industry leaders are working to better understand how data collection may not effectively document what people with disabilities need. They're improving the process by doing a better job of gathering information.

Image shows a white puzzle over red background; the background is visible through missing puzzle pieces.
When implementing disability-inclusive practices, governments should consider several key factors, as illustrated through a federal request for information (RFI), and in a new local data standard.

Governments are focused on accessibility for numerous reasons, one being the new accessibility mandate for digital government content. However, access can and should go beyond compliance; and as artificial intelligence continues advancing, high-quality data should be the foundation.

The White House Office of Science and Technology Policy (OSTP)’s May RFI is a catalyst for this work, demonstrating federal action in disability equity. Through the RFI, OSTP aims to collect input to guide development of the Federal Evidence Agenda on Disability Equity. The RFI’s intent is to increase the federal government’s ability to advance equity, through data-informed policy decisions.

One group that submitted comments on the RFI is the Center for Democracy and Technology (CDT). People who are disabled face the risk of discriminatory outcomes through algorithmic bias, its Policy Counsel for Disability Rights in Technology Policy Ariana Aboulafia said. CDT’s goal in commenting was to make recommendations on creating inclusive and representative data sets about disability that can be used as inputs to reduce that risk.

“Algorithmic outputs are created as a result of inputs, and those inputs come from data sets,” said Aboulafia, who recently authored a CDT report assessing problems and potential solutions in disability data equity.

DISABILITY DATA GAPS


People with disabilities may be underrepresented in data sets, she said, because data collection may not acknowledge the multiple definitions of disability — resulting in inaccurate, incomplete data sets. Disability is often defined by its legal definition, from the Americans with Disabilities Act, but it can also be viewed through medical, social and identity-based models.

There are different models of disability, Aboulafia said, but there are different kinds of disabilities, too: “There are just as many ways to be disabled as there are people with disabilities.”

Some disabled identities are more stigmatized than others, she said, which may impact whether people identify as being disabled. Culture and location play a part, too; and people with disabilities are disproportionately incarcerated or institutionalized, making them harder for data collectors to represent.

RISKS OF NONREPRESENTATIVE DATA


If data collectors are missing these populations and data sets do not represent them, budget allocations may miss them — and risks like algorithmic bias may creep in. Hiring tools, Aboulafia said, may have disproportionately negative outcomes for people with disabilities.

The tools, she explained, are trained on pattern recognition — trained to recognize patterns in traits like eye contact and vocal cadence, based on the average. People with disabilities may have different vocal cadence or eye movements that could cause them to be screened out of a job, just for having what a system recognizes as atypical traits. This is because these tools measure based on their initial data set, which often is not entirely disability-inclusive.

BEST PRACTICES FOR INCLUSIVE DATA


Designing a system’s input on the front end can save teams from having to mitigate discrimination later on the back end, Aboulafia said. Being more inclusive for people with disabilities actually makes things better for people without disabilities, too, she added.

Inclusive data collection doesn’t only define disability, Aboulafia said. It involves being transparent with respondents about how their responses will be used, having data protection measures to ensure their privacy, and being forthcoming about the potential real-life impact of their responses.

Surveys themselves should use plain language to be more accessible, she said, and use assistive technology like screen readers to be usable by all. Accessible survey formats should be used by any government entity looking to create representative data sets, Aboulafia said.

INCLUSIVE DATA FOR LOCALS


The city of Boston has been working to implement equitable data collection practices, including making a change last year to address data gaps in marriage licensing for people who would rather file without a gender option.

Building on this, the city announced the creation of a Disability Data Standard in July. The initiative is a collaborative effort led by its Disabilities Commission, working with the Department of Innovation and Technology. It offers guidelines for officials to better understand when and how to collect information about constituents’ disabilities.

The standard also highlights transparency about why data is being collected and how it will be used and kept private as best practices, which aligns with Aboulafia’s insight.

According to Mariangely Solis Cervera, city chief of equity and inclusion, the mayor’s vision is to make Boston a home for everyone, both at a policy level and in terms of day-to-day experience. To address constituents’ experience, she said it was important to better understand the lived experience of people with disabilities.

A key piece of this new standard is asking residents outright which of 11 access and functional needs options they require to access services and events, or letting them write their own response.

A second phase of this work is underway, Solis Cervera said. Officials are exploring how Boston can better support people with disabilities who are on staff; and externally, how they can ensure voters have the necessary access.

As detailed in the standard’s first guideline, disability data collection should be part of an entity’s broader accessibility strategy. The city’s Chief Digital Officer Julia Gutiérrez noted that while inclusive data collection can inform accessible service delivery, data collection on its own is not sufficient: “It’s what we’re doing with that information.”

Creating the standard involved collaboration between city teams, but also recruiting people for focus groups who had diverse lived experiences in their age, race, ethnicity, gender and type of disability, Gutiérrez said. The process built on existing research but also examined unique experiences and needs in the community.

Boston provides more than 550 city services, not all of which are currently digital, so this initiative will be ongoing, Solis Cervera said, and will require organizational change, relationship- and trust-building, and accountability: “It takes a lot of courage, which is why we’re not going to end here.”
Julia Edinger is a staff writer for Government Technology. She has a bachelor's degree in English from the University of Toledo and has since worked in publishing and media. She's currently located in Southern California.