Somewhere in a funder's annual report, there is a photograph of a community meeting. Twelve people around a table, one flipchart, a facilitator with a marker. The caption says the project "engaged the community." Nobody in that room would describe what happened the same way, and that gap — between what gets reported and what actually occurred — is the whole problem with inclusive community engagement metrics for research partnerships.
I have sat on both sides of that table. I've been the researcher writing the progress report, and I've been the person in the community asking why the "consultation" happened after the budget was already spent. The metrics I used in the beginning were, frankly, embarrassing: number of meetings held, number of attendees, number of flyers distributed. Counting bodies. That's not engagement measurement, that's event logistics.
Here's what changed for me, and what this article is about: if you want metrics that actually capture inclusion in a research partnership, you have to measure decision rights, not attendance. Everything else is decoration.
Key Takeaways
- Attendance is not inclusion. A room full of people with no influence over the agenda is a performance of engagement, not an instance of it.
- Track power metrics (who decides, who signs, who is named) alongside participation metrics.
- Community partner retention over the full project lifecycle is one of the most honest indicators you have — and one of the least reported.
- Build metrics with partners before the project starts, not after the first reporting deadline.
- Report the uncomfortable numbers. A partnership that reports only successes is telling you something about itself.
Inclusive community engagement metrics that measure power, not headcounts
The reason most engagement dashboards are useless is structural: they were designed by the institution, for the institution, to satisfy an institutional reporting requirement. They measure what is easy to count from the inside.
I watched this play out on a health study where the community advisory board met eleven times over fourteen months. Eleven meetings. Excellent number to report. What the report didn't say: the board's recommendations on recruitment were overruled twice, and the compensation for members' time was delayed by four months because procurement had no category for it. The engagement metric was perfect. The engagement was not.
The three layers of measurement
What works, in my experience, is a layered structure. Each layer answers a different question, and you need all of them because none is sufficient alone.
- Reach — who is in the room, and does that composition match the population the research claims to serve? This is where demographic breakdown, language of materials, and accessibility accommodations live.
- Influence — what changed because partners were involved? Concrete: agenda items they set, protocol amendments they triggered, questions they rewrote, dissemination channels they chose.
- Relationship — is the partnership durable? Retention rates, response times, whether partners return for the next project, whether anyone gets paid on time.
Most programmes measure reach and call it a day. Reach is the easiest layer and the least informative. A partner who shows up to every meeting and is never listened to will score perfectly on reach.
Why decision rights belong in your indicator set
Ask a simple question about any research partnership: which decisions require the community partner's sign-off? Not consultation. Sign-off.
If the honest answer is "none," you have measured everything and learned nothing. Decision rights are specific and auditable — budget lines they control, hiring input they hold, authorship positions they occupy, data governance roles they fill. Each of these is countable, and each tells you more about inclusion than a hundred attendance sheets.
How do you put numbers on inclusion without faking it?
You build the indicators with the partners, and you accept that some of the most important ones will be uncomfortable to report.
Early on, I made the mistake of designing the metric set myself and presenting it for "feedback." The feedback was polite and minimal. Of course it was — I had handed people a finished artefact and asked them to critique my homework. What actually worked, on a later project, was a half-day session where we agreed on five indicators together. Took four hours. Changed the entire tone of the partnership, because partners could see their own priorities in the reporting template.
A worked example of an indicator set
Here is a structure I've used since, adapted per project. It is not a standard, it is a starting point for negotiation — which is the point.
| Layer | Indicator | How you compute it |
|---|---|---|
| Reach | Representation gap | Share of engaged participants from each priority group, compared with that group's share of the affected population |
| Reach | Access completion | Proportion of engagement sessions held with interpretation, translated materials, or accessibility support where requested |
| Influence | Co-decision rate | Number of project decisions where a community partner held formal sign-off, divided by total major decisions recorded |
| Influence | Uptake of partner recommendations | Recommendations formally logged, and the share adopted, modified, or rejected with written reasons |
| Relationship | Partner retention | Share of community partners still active at project close, measured against those active at month three |
| Relationship | Payment timeliness | Median days between invoice or claim submission and payment to community partners |
The payment timeliness one is not a joke. I have seen it do more damage to trust than any methodological disagreement. Partners who wait months for modest compensation learn exactly how much their time is valued, whatever the meeting minutes say.
Setting thresholds instead of chasing perfection
Absolute scores are meaningless in isolation. What matters is whether you agreed, in advance, on a minimum acceptable level and what happens if you fall below it.
A worked pattern: if co-decision rate falls under roughly a third of major decisions across two consecutive reporting periods, the governance arrangement gets reviewed with partners before the next phase begins. That's it. A number, a trigger, a consequence. Without the consequence, you've built a dashboard nobody is obliged to look at.
Which tools and frameworks are worth using?
There is no shortage of instruments. There is a shortage of instruments that survive contact with a real budget and a real timeline.
The published landscape includes several named assessment tools — collaborative partnership scales, coalition self-assessments, engagement indices designed for research settings. They are useful as conversation structures, less useful as scoring machines. I've tried running a full partnership assessment instrument mid-project and got a 40-page output that nobody read. We then spent twenty minutes in a room and reached the same conclusions.
When a structured instrument earns its keep
Two situations justify the overhead. First, multi-site projects where you need comparability across sites and cannot rely on a shared informal read. Second, funder-mandated reporting where a recognised instrument lends legitimacy to findings the partnership already knows.
Outside those, a short custom set built with partners beats a comprehensive borrowed one. Real talk: the value of the assessment is in the discussion it forces, not the score it produces.
What about the community engagement index?
Index-style approaches — rolling multiple indicators into a single composite score — are tempting because they compress well into a report. I'd use them with caution. A composite hides exactly the information you need: a project can score well overall while performing badly on the one dimension that matters most to a specific community. Report the components. Let readers see the unevenness.
What does inclusive engagement look like when it actually works?
A project I worked alongside had a rule I initially thought was excessive: no research question entered the protocol until a community partner had signed off on its wording. Not consulted on. Signed off on.
The consequence was measurable. Recruitment ran ahead of the original projection, and the dropout rate among participants from the priority population stayed low. Was the sign-off rule the cause? I can't prove it, and I won't pretend a clean causal claim. But the questions were phrased in language people recognised, and the people who recognised it were the ones who wrote the phrasing. Draw your own conclusion.
The metrics nobody wants to report
If your engagement reporting contains only positive indicators, that is a finding about your reporting, not about your engagement. Include the failures. Partners who left, and why. Recommendations rejected, and on what grounds. Meetings cancelled for lack of quorum. Complaints received and how they were resolved.
A progress report I once reviewed listed a community partner as "engaged throughout." That partner had formally withdrawn five months earlier. The engagement log was never updated. Which is why I now insist that retention and withdrawal are reported as explicit figures rather than inferred from a list of names.
Where to start
Pick one decision in your next project that currently belongs to the research team alone. Hand its sign-off to a community partner. Then measure what happens — including if it goes badly, because the first time often does, and the record of that first time is the most valuable data point you'll collect.
The metrics follow from the arrangement, not the other way around. You cannot measure inclusion into existence with a good dashboard; you can only describe it honestly once it's there. And the honest description, awkward numbers included, is what makes the next partnership possible.
