Ripple Effects Mapping: capturing the effects nobody planned

REM is a participatory group method for retrospectively identifying and mapping the effects of a programme, including secondary, indirect and unintended ones. It is built for community-level work — Extension programmes, coalitions, local development initiatives — where effects propagate through relationships that no monitoring system observes.

Its central premise is that the knowledge is distributed. No single participant knows what a programme produced; each knows a fragment, and the fragments connect. The session is designed so that one person's account prompts another's, and the map grows by association rather than by interrogation.

The typical REM session combines appreciative inquiry interviewing in pairs, group mind-mapping of the effects, and facilitated in-depth probing to push past the first, obvious layer. The output is a visual map, subsequently coded and analysed, plus a transcript of the discussion that produced it.

REM emerged from US Cooperative Extension evaluation in the early 2010s, developed by Scott Chazdon, Mary Emery, Debra Hansen, Lorie Higgins, Rebecca Sero and colleagues. Kollock and co-authors introduced it in Ripple Effect Mapping: A "Radiant" Way to Capture Program Impacts, and Emery and colleagues set out the practical variants in Using Ripple Effect Mapping to Evaluate Program Impact.

It was designed against a specific frustration: Extension programmes were producing visible community change that their reporting could not capture, because the change ran through volunteers, informal networks and follow-on activity that nobody was funded to track.

The method's claim to evaluate unintended outcomes specifically was tested by Peterson and colleagues in Evaluating unintended program outcomes through Ripple Effects Mapping, and a fuller methodological treatment appears in the SAGE Handbook of Participatory Research and Inquiry. A worked application is described in Participatory Evaluation and Learning.

A three-year rural food-systems programme funded a farmers' market, a school garden and a series of cooking workshops. Its reporting covered market vendor numbers, garden beds built and workshop attendance.

The REM session brought together 18 people: vendors, teachers, two parents, a public-health nurse, the market manager, and the owner of a café who had never been part of the programme at all.

The first mapping layer reproduced the reporting. The second layer began to diverge: two vendors had started supplying the café, which was the reason its owner was in the room. The school garden had generated a parent volunteer group that had subsequently taken on a separate playground project with no programme involvement. The nurse had begun referring patients to the cooking workshops, which had not been anyone's plan.

The third layer produced the finding that mattered. The parent volunteer group had absorbed the two most active volunteers from the town's existing community centre, which had cancelled its own after-school programme. That is a real, negative, second-order effect, and it emerged only because the facilitator asked directly what had stopped happening. It would not have appeared in any monitoring system, and it was the single most useful thing the evaluation produced.

• Stopping at the first layer. Without persistent "and then what?" probing, REM produces an activity map at considerably more expense than a report would have cost.
• A room of insiders. Staff and committed participants know the intended effects. The ripples live with people at the edge of the programme.
• Appreciative framing taken as prohibition. Appreciative inquiry is a way in, not a rule that only good news may be mapped. A map with no negative branches has been facilitated badly.
• No verification. Group recall is confident and sometimes wrong, particularly about dates and about who did what first. Significant claims should be checked afterwards.
• Treating the map as the deliverable. A photograph of a wall of sticky notes is not an analysis. Coding and chain-tracing is where the finding is.
• Attribution by adjacency. Appearing on the map is not evidence the programme caused it. Distance from the centre is a claim about connection, not about cause.

A defensible REM has: a participant list documented with the rationale for who was invited; branches reaching at least the third order; negative and terminated effects present; a coding scheme stated in advance; and follow-up verification of the claims the evaluation intends to rely on.

The clearest signal of a poor session is a map that reproduces the programme's logic model. If everything on the wall was already in the plan, the session has recorded intentions rather than effects.

REM shows connection, not causation. A chain on the map means participants believe one thing followed from another; establishing that it did requires a different method, and REM findings are frequently over-claimed on this point.

It is also retrospective and reconstructive. Group memory compresses time, misattributes initiative and favours the memorable, and the social dynamic of a room can suppress dissenting accounts — the person who thinks the market damaged the existing grocery may not say so in front of the market manager.

It says nothing about magnitude. Twenty branches may represent twenty small things. And it does not scale: REM is a community-level instrument, and applying it to a national programme produces abstraction rather than ripples.

Outcome harvesting covers similar ground — unplanned outcomes, retrospectively identified — but formulates them as individually verifiable statements and substantiates them independently, which is the discipline REM lacks.

Most Significant Change works through individual narratives and stakeholder selection rather than collective mapping.

Participatory evaluation is the broader family; REM is one session format within it.

Case-based impact evaluation is the follow-on when a particular ripple turns out to matter enough to establish properly.

The map is only the halfway point. The analysable material is the session transcript plus the digitised map, and the coding — against community capitals or against the theory of change — is ordinary qualitative analysis at a scale that rewards doing it systematically.

Where a programme runs REM in several communities, the comparison across maps is where the real finding usually sits: which ripple types recur, which are site-specific, which branches consistently die at the second order. Evidano supports Ripple Effects Mapping as a named methodology on the transcript and coding side; building the map is a facilitated human activity and should stay one.