The Evidence Audit Checklist

Purpose: A five-minute self-review to help you decide whether a single outcome statement is ready to share externally — before it reaches a board, funder, or public report.

How to use this (30 seconds)

Pick one outcome statement you plan to report — a single sentence you intend to say out loud or put in writing. For example: "Our program improved participants' wellbeing."Run that one statement through the five sections below, ticking each box you can honestly answer "yes" to. Then use the scoring guide to place your statement into one of four states, and note the single next evidence action. Repeat for each statement that matters.

Section A — Is the evidence actually there?

I can name the specific source of evidence behind this statement (not a general impression).
The evidence is recent enough to still be true for the period I'm reporting on.
The evidence is reasonably complete — not one or two responses standing in for everyone.
I know who collected the evidence and how, and I'm comfortable explaining that.

Section B — Does the evidence support this statement?

The evidence is actually about the thing this statement claims (not a near-neighbour topic).
The evidence is strong enough to carry the weight of the statement, not just suggestive of it.
The sample or scope matches the group I'm making the statement about.
I've considered whether something other than our work could explain the result.

Section C — Is the language matched to the evidence?

The statement avoids over-claiming words the evidence doesn't earn (e.g. "transformed", "eliminated", "proven").
Where I've implied our work caused a change, the evidence actually supports cause — not just that two things moved together.
I've hedged appropriately ("early signs", "for those surveyed", "reported") where the evidence is partial.
The statement would read the same to a sceptical reader as it does to me.

Section D — Are the limits and gaps visible?

I've named what's missing or still uncertain, rather than letting the statement imply full certainty.
I haven't quietly left out results that point the other way.
Any important reporting limits (small sample, short timeframe, self-reported data) are stated near the claim, not buried.

Section E — Can you trace it?

If someone asked "how do you know?", I could link this statement back to its source evidence.
That source trail is something I could produce on request, not reconstruct from memory.
Another person in my organisation could follow the same trail and reach the same statement.

Scoring & interpretation

Don't add up a score. Look at which sections your "no" answers landed in. Read the rules in order and stop at the first one that matches:

Not yet ready— Any "no" in Section A (evidence there?), Section B (does the evidence support the statement?), or Section E (can you trace it?). The foundation is missing. → Fix the evidence, its fit to the claim, or its source trail first. Don't report this statement yet.
Internal only— Any item you genuinely can't answer yet. → Keep it for internal discussion until someone can review it. Don't put it in an external report.
Cautious (share with a stated limit)— "No" answers only in Section D (limits/gaps visible?). → Add the reporting limit next to the claim, then share.
Ready (after softening)— "No" answers only in Section C (language matched?), with A, B, D and E clear. → Soften the wording to match the evidence, then it's ready.
Ready to share — All boxes ticked. → Share it as written.

Evidence gaps set the ceiling; language can't lift it. A "no" in Section A, B, or E caps a statement at Not yet ready, no matter how good the wording is.

Your next evidence action

Look at the first section where you answered "no", working from A to E. Fixing the earliest gap is almost always the single most useful thing you can do next — because later checks depend on it. Write that one action down before you move on.

Where CIIS fits

This checklist walks one statement through one review. CIIS (Collaborative Impact Intelligence System) does the same evidence review systematically, across a whole report — surfacing what can be said, what must be said cautiously, what cannot yet be said, which evidence gaps remain, and the source trail behind each statement. It's governance-first reporting: evidence surfaces, people decide, and limits stay visible. If that's the kind of assurance you want before your next report goes out, we'd welcome a conversation about a pilot.

Evidence surfaces. People decide. Limits stay visible.

Request a pilot discussion →

Next Steps

If this topic resonates with challenges you're facing, consider:

Request a pilot discussion ciis trust how it works