# Researching with people with disabilities: recruiting, accommodating, and running sessions well

> How to recruit and run sessions with disabled participants: profiles by assistive technology, accommodations, and why a test cannot certify accessibility.

*Tags: ux, research, accessibility, recruiting, inclusion*

---


> [!info] Quick Definition
> Researching with people with disabilities means including them as real participants in usability studies, not only as the people a design is meant for. Doing it well requires recruiting by how people use technology, preparing access and accommodations before the session, and understanding that a test reveals problems but does not certify accessibility on its own.


## Why is this research different?

Because people use the product differently. Someone who navigates with a screen reader, a keyboard, or voice control runs into obstacles that a typical user never sees. General usability findings can miss those barriers, and accessibility findings can be lost if nobody who uses those technologies took part.

The W3C frames it this way: evaluating with people with disabilities and with older users identifies usability problems that conformance evaluation does not detect on its own.

## How do you choose participants?

Section508.gov recommends recruiting by functional capabilities and assistive technology use, not by diagnosis. Useful questions cover which devices the person uses, which assistive technologies they rely on, and how comfortable they are with them. Avoid keeping lists of people by condition.

The profile depends on the product. The W3C gives an example: for an internal app used by accountants, advanced assistive technology users may be the right fit, while for a public benefits application, novice users are more representative. A single participant does not represent their whole group, so the scope of the study should be reported clearly.

Smashing Magazine suggests starting small: include one or two participants with disabilities as part of a larger group of five to ten. Section508.gov, for its part, speaks of six to eight participants in one-on-one studies.

## Where do you recruit?

A general recruitment panel rarely reaches these profiles. Section508.gov and Smashing Magazine mention several paths:

- **Associations and community organizations**, which often keep volunteer or research lists.
- **Local chapters** of organizations that serve people with disabilities, such as a local chapter of the National Federation of the Blind, which Smashing Magazine mentions.
- **Panels from companies specialized in accessibility**, if you first ask whether they have experience with this kind of participant.
- **Employee resource groups** inside your company, where they exist.
- **Direct announcements** in relevant communities, with a clear description of the study and of the compensation.

Section508.gov also advises against buying customer lists for recruitment. Ask vendors to spread the call instead of obtaining data about people who have not consented.

## What accommodations should you prepare?

Ask each person beforehand what they need. Section508.gov and Smashing Magazine mention:

| Accommodation | When it helps |
|---|---|
| Accessible venue or remote session | When travel is hard or the person prefers their own device |
| Transport or reimbursement | So distance does not prevent participation; Smashing Magazine suggests adding $25 to $40 to the incentive |
| Sign language interpreters | For Deaf participants; Section508.gov suggests two if the session runs longer than 60 minutes |
| Real-time captioning (CART) | For Deaf or hard-of-hearing participants, for live multimedia |
| Alternative formats | Consent forms and materials readable by screen readers |
| Personal assistant | If the person needs one; Section508.gov asks you to reimburse the cost separately from the incentive |

Smashing Magazine recommends asking participants to bring their own equipment and having a backup plan with test technology, and leaving an hour between sessions for setup. Section508.gov suggests over-recruiting by 15 to 25% to account for no-shows.

## How should the session run?

Section508.gov proposes sessions of about 90 minutes, including setup and breaks, and setting aside time at the start for consent forms. Smashing Magazine adds preparation details: email consent forms in a format a screen reader can read, avoid relying on PDFs, do not use color as the only cue, and, for participants with cognitive disabilities, give instructions in more than one format and break tasks down.

The venue matters too: accessible routes with elevators, adjustable tables, nearby restrooms, and an evacuation plan. Confirm technical requirements, such as ports or connectors, before the session.

A good practice: before the test, fix the basic problems a conformance review finds, such as missing alt text or insufficient contrast. That way the participant can focus on the flow, not on barriers you already knew about.

## When is a user test not enough?

The W3C warns that involving people with disabilities has many benefits, but on its own it cannot determine whether a site is accessible. It must be combined with an evaluation against WCAG. Moreover, results from only a couple of people cannot be generalized, and small studies are useful for finding problems but are not robust enough for statistical conclusions.

If your study aims to decide a usage trend, a small sample is not enough. If it aims to find real barriers, it is the most direct way to do so.

## Example

Illustrative example: a team is redesigning the application flow for a public service. Before testing, it reviews conformance and fixes two basic problems: buttons without accessible names and a form with labels that were not visible. It recruits three screen reader users and two keyboard-only users through a local association. It offers in-person sessions at an accessible venue or remote sessions, depending on each person's preference, and asks beforehand what support they need. They find that the error message on the date field is not announced to the screen reader, a problem the automated check had not flagged. The team fixes it and retests with two participants.

## Common mistakes

- **Recruiting by diagnosis** instead of by technology use.
- **Not asking about accommodations** and discovering them on the day of the session.
- **Using an inaccessible venue**, which excludes participants before they start.
- **Relying on a single test** as an accessibility certification.
- **Forgetting compensation and transport**, which can prevent participation.
- **Generalizing one person's finding** to a whole group.

For what comes after the session, [accessibility annotations](/en/wiki/artifacts/accessibility-annotations/) help communicate findings to developers, and [screen reader testing](/en/wiki/techniques/screen-reader-testing/) describes one specific technique. If you need to widen the criteria beyond the visual, [[cognitive-accessibility|cognitive accessibility]] explains how to include memory and attention in design. And for general recruiting, [usability testing](/en/wiki/research/usability-testing/) describes the base method.

## Sources

- [Involving Users in Evaluating Web Accessibility, W3C WAI](https://www.w3.org/WAI/test-evaluate/involving-users)
- [Conducting User Research with People with Disabilities, Section508.gov](https://section508.gov/develop/usability-testing-with-people-with-disabilities)
- [How To Conduct UX Research With Participants With Disabilities, Smashing Magazine](https://www.smashingmagazine.com/2018/03/tips-conducting-usability-studies-participants-disabilities/)


---

Source: https://www.fernandoux.com/en/wiki/research/research-with-disabled-participants/
