Last spring, a research team I was advising held its fourth "co-design" workshop with a group of disabled self-advocates. The academics came in with a pre-filled protocol. The community partners came in with two years of frustration. By the third hour, one participant said the sentence that has stayed with me since: "You keep asking us how to design the study, but you've already decided what matters. We're just decoration."
That is the fault line running through inclusive research co-design with disability community partners. The vocabulary has spread fast. The practice has not caught up. I've worked on the applying side, the advising side, and the awkward middle where a funder wants "meaningful engagement" but the timeline gives you six weeks. Here is what I've learned about making it real, and where teams keep getting it wrong.
Key Takeaways
- Co-design is one slice of co-production: shared control over the questions, tools, and interpretation, not just the feedback session.
- Power imbalances are structural, not personal. Fixing them takes budget, accessible logistics, and a willingness to lose control of the agenda.
- Compensation, accessible formats, and clear decision rules are non-negotiable; skipping them produces the exact tokenism you're trying to avoid.
- The most common failure: recruiting partners after the research question is locked.
Inclusive research co-design: what it actually requires from you
The word "co-design" has drifted. In a lot of funding proposals I've reviewed, it means "we ran a focus group and told them the plan." In honest usage, co-design means disabled people and researchers share authority over the design decisions: what gets studied, how, with which tools, and how the findings get read.
Here's the thing that trips teams up. Co-design is a component of co-production, not a synonym. Co-production is the whole arc — planning, doing, reviewing. Co-design is the shaping phase. You can have co-design without co-production (community partners shape the interview guide, then vanish while you write the paper). That's where most projects sit, and that's precisely where the goodwill curdles.
The definition problem nobody wants to name
Consultation is when you ask. Co-design is when they decide with you. Co-production is when you share the whole house.
I use this distinction in every kickoff meeting now, on a whiteboard, in plain language. It sounds pedantic. It saves months. When a funder says "co-designed," they often mean "we consulted." When a community partner says "co-designed," they usually mean "you didn't override us when it mattered." Those are different projects wearing the same name.
In my opinion, the fastest way to fracture trust is to keep using the co- word while running a consult- process.
The power problem costs money, and most grants don't cover it
Researchers talk about power-sharing like it's a mindset. It's a budget line. Accessible venues. Sign language interpretation. Captioning. Travel reimbursement for people whose disability makes an unpaid afternoon at a university genuinely expensive. Compensation for time, at a real rate, not a gift card.
When I helped a small team design a co-analysis workshop last autumn, the accessibility logistics ate roughly a quarter of the engagement budget — before anyone had been paid. Renegotiating the budget mid-project to cover that is a conversation almost no one wants to have with their institution. I've watched teams decide against proper accommodation because the money was already committed. That's not a values failure at the individual level. It's a structural one, and it produces the same result.
Who does the work, and who gets credited?
Ask yourself, plainly: whose name is on the paper? Whose voice is in the methods section? Who gets invited to the conference?
Real talk: co-design without credited authorship is unpaid labor with better branding. The fix isn't complicated. Agree on it in writing at the start — contribution, credit, and what happens if someone needs to step back. Disabled partners often carry more admin and health load than a research calendar assumes. Build in exits that don't look like failure.
- Authorship criteria agreed before the first workshop
- A named budget line for compensation and access costs
- Documented decision rules: what do we decide together, what do I own, what does the funder own?
- An explicit right to say no without consequence
What shared control looks like in practice
Look, most failures happen before the first meeting. If the research question is already written into the grant, your co-design is theater. The decision you've reserved is the one that matters most.
The projects I've seen work move the question-setting into the room. That means the protocol draft is a draft, printed in large print, with space to cross things out. It means a partner can strike an entire aim. It means the team has practiced saying "we don't know yet" out loud.
A comparison that keeps me honest
| Dimension | Consultation | Co-design | Co-production |
|---|---|---|---|
| Who sets the question | Researchers | Shared | Shared throughout |
| Decision power | Advisory only | Joint on design choices | Joint on design, delivery, interpretation |
| Data tools | Presented for feedback | Built together | Built and tested together |
| Analysis and writing | Researchers | Usually researchers | Shared, with credit |
| Typical risk | Tokenism | Design shared, power retained elsewhere | Higher coordination cost |
Use the middle column as your honest checkpoint. If your project sits there and the claim is "co-produced," say so. Precision isn't pedantry when the alternative is misrepresenting the people in the room.
How do you avoid tokenism?
Tokenism is when a disabled partner's presence is used to legitimize a decision they didn't influence. The tell is a partner who is invited, thanked, and never asked to change anything.
Three practical moves I've used. First, give the group something real to decide during the first meeting — even small — so the tone is set immediately. Second, reserve at least one recurring meeting slot where the agenda is set by partners, not researchers. Third, when a community recommendation is overruled, say so in writing to the group and explain why. That last one is uncomfortable and it's the one that builds trust, because it proves you're not performing.
Co-design vs consultation vs co-production: a clean framework
Map it on two axes: who decides and what is subject to decision. Consultation decides what "would be nice"; co-design decides what gets built; co-production decides what gets said. If you can't name the decisions your partners actually hold, you know your answer.
How should community partners be compensated?
Pay for time the way you'd pay a consultant — because that's what they are. Cover the access costs separately and in advance: interpretation, accessible transport, captioning of recordings. Offer flexible formats (written, spoken, video, in-person, async) and let people choose their own. And crucially, agree in advance on what happens when someone's capacity changes. Chronic illness doesn't follow your Gantt chart.
I made a mistake early on: I assumed one shared workshop format would suit everyone. It suited the researchers. Two partners disengaged quietly over a month, and I only noticed at the evaluation. Async options and a short check-in after each session would have caught it.
What does inclusive research co-design actually produce?
Better instruments, for one. In the project I advised, partners rewrote two survey items that were unintentionally ableist in wording. Those items had already passed ethical review. They would have produced skewed answers the team would have confidently misread. The change cost an afternoon. The cost of not doing it would have been a published paper based on bad questions.
Getting started without a blueprint
There isn't a template that travels well. Each disability community, each team, each institution carries different friction. What travels is the order of operations: budget the access and compensation first, agree on decision rights before the first session, share the question-setting, and be explicit about credit.
The discomfort is the whole point. The moment co-design stops feeling slightly threatening to the researchers is usually the moment it has become decoration. The question that stays with me, long after that fourth workshop, is simple: if a disabled partner could veto one thing you've already decided, what would it be — and would you let them?
