Participatory and Empowerment Evaluation

One family, two intentions

“Participatory evaluation” names a family, not a method, and the family’s most useful map remains Cousins and Whitmore’s: two streams with different justifications, sharing techniques but not purposes [1].

Practical participatory evaluation (P-PE) involves programme staff and other stakeholders in the evaluation because participation produces better evaluations — more relevant questions, better-grounded interpretation, findings that the organisation actually uses. Its warrant is utilisation, and its intellectual kinship with utilization-focused evaluation is close: participation is the means; decision support is the end [1].

Transformative participatory evaluation (T-PE) inverts the justification. Rooted in participatory action research traditions from the global South, it treats participation itself as an exercise of power: the point is that people normally treated as data sources become producers of knowledge about their own situation. Better information is welcome; redistributed control is the goal [1].

The distinction matters because the two streams fail differently. A P-PE that excludes the community can still succeed on its own terms; a T-PE that leaves control with the commissioning organisation has failed on its own terms no matter how good the report is.

Three dimensions, one honest picture

Cousins and Whitmore locate any participatory design on three dimensions of process [1]:

1. Control of the evaluation process — from evaluator-controlled through shared to community-controlled.
2. Stakeholder selection — from primary users only to all legitimate groups, including those the programme is done to rather than with.
3. Depth of participation — from being consulted, through contributing to specific tasks, to involvement in all phases: framing questions, collecting data, analysing, interpreting, reporting.

The dimensions are independent, and most real designs are uneven across them — deep participation in data collection with no participation in analysis is the sector’s default pattern, and it is the shallowest of the three to concede, since interpretation is where meaning is made. An honest design statement says where the evaluation sits on each dimension, phase by phase.

Empowerment evaluation

At the community-control end of the spectrum sits empowerment evaluation, launched by David Fetterman in his 1993 American Evaluation Association presidential address as “the use of evaluation concepts, techniques, and findings to foster improvement and self-determination” [2]. The programme’s own community conducts the evaluation; the evaluator’s role shifts to coach and critical friend — teaching evaluation logic, facilitating the group’s self-assessment, and pushing back on comfortable conclusions without owning the verdict [2][3].

As practised and documented since, the approach runs on a compact core: the group articulates its mission, takes stock of its current activities against that mission — typically by rating and evidencing them — and plans for the future with goals and strategies it will itself monitor. A set of published principles governs the practice, with community ownership, inclusion, democratic participation, capacity building and accountability among them [3]. Two commitments distinguish it from generic participation: evaluation capacity is meant to remain in the community as a lasting asset, and self-determination is an outcome the evaluation is accountable for, not a by-product.

The rigour debate — presented, not smoothed over

Empowerment evaluation has been controversial since its founding statement, and the disagreement is substantive. This site’s editorial rule is to present live methodological divides with both sides cited; this is one of them.

The critics’ case. The challenge arrived almost immediately — most prominently from Daniel Stufflebeam in the mid-1990s exchanges in Evaluation Practice — and has been renewed since. It has two prongs. The first is definitional: Patton, himself no enemy of use-driven evaluation, pressed the question of whether empowerment evaluation is actually distinguishable from the collaborative and participatory approaches that surround it — if the distinctive element is an advocacy commitment rather than a distinct methodology, the label may claim more than the practice delivers [4]. The second is epistemic: Nick Smith’s analysis argues that empowerment evaluation is best understood as an ideology and social movement rather than a bounded evaluation model — that when a group evaluates itself with a facilitator committed to the group’s empowerment, the arrangement structurally discourages negative findings, and calling the result “evaluation” trades on a warrant (independent judgement of merit) that the practice has given away [5].

The proponents’ case. Fetterman and Wandersman’s reply reviews the approach’s record and answers the critics directly [6]. Self-assessment is not the absence of standards but a different discipline: ratings must be argued and evidenced within the group, the critical friend exists precisely to challenge self-serving conclusions, and accountability is one of the approach’s stated principles, not an omission [3][6]. On validity, the proponents’ strongest ground is local knowledge: people inside a programme know where the bodies are buried, and processes they own surface problems that external evaluators are routinely lied to about. And on use, the empowerment claim inherits the utilisation literature’s oldest finding — evaluations imposed from outside are the ones most often ignored.

Where this leaves a commissioner. The two sides are optimising different guarantees, and the honest resolution is fit-for-purpose, declared in advance. Where the requirement is an independent account that can stand against interested parties — funding gates, compliance, contested claims — the critics’ objection is decisive, and an independent design should be used. Where the requirement is a community or organisation building durable capacity to steer its own work, empowerment evaluation is doing something an external judgement cannot, and should be labelled as what it is rather than dressed as independent evaluation. Mislabeling in either direction is the actual malpractice.

The methods toolbox

The family’s techniques are shared across the spectrum, inherited in large part from participatory rural appraisal traditions: community mapping and ranking exercises, seasonal calendars, community scorecards in which service users rate services and negotiate improvements with providers, participatory data collection by community members, and structured participatory analysis sessions in which community participants — not the evaluator — sort, interpret and challenge the evidence [1][3]. Story-based techniques with participatory selection, notably Most Significant Change, sit naturally in this toolbox. What makes any of these participatory rather than extractive is not the technique but the control settings around it — the three dimensions above.

Who sets questions, who interprets, and who owns the data are also accountability questions, with obligations that exist independently of any approach choice; they are treated in accountability to affected populations. Participatory designs centring power relations specifically around gender connect to feminist and gender-responsive evaluation.

Honest tests for tokenism

Participation is the sector’s most inflated claim, so the tests should be blunt. An evaluation’s participation claim fails when:

• The questions were fixed before anyone participated. Community input to pre-set questions is consultation, and should be called that.
• Participation stops at data collection. Enumeration by community members is employment, not participation, if analysis and interpretation stay with the evaluator.
• The selection of participants was made for convenience. A committee of programme-friendly notables is not “the community”; who was in the room is reportable information.
• Nothing could have been changed. If no finding could have altered the programme, participants’ time was spent legitimising a decision already made.
• The data left and never came back. Ownership shows at the end: whether participants can access, contest and use the results [3].

Publishing the answers to these five tests in the evaluation report — a paragraph, not a chapter — costs little and converts a rhetorical claim into an inspectable one.