Participatory Evaluation: stakeholders as evaluators, not data sources

The canonical map is Cousins and Whitmore’s Framing participatory evaluation, which distinguishes practical participatory evaluation — stakeholder involvement in service of evaluation use and organisational decision-making — from transformative participatory evaluation, where the point is shifting power: building marginalised groups’ control over what counts as evidence about their lives.

The same paper supplies the three dimensions along which any participatory design can be honestly described: who controls the evaluation’s technical decisions, whose selection — how broad the stakeholder range is, and depth — whether involvement runs from consultation to full co-inquiry across all stages.

Locating a design on those dimensions, in writing, before fieldwork is the single best predictor of a coherent process: it converts “participatory” from a mood into a specification.

• Questions: stakeholders surface the questions the programme’s theory forgot — often about burden, dignity, and side effects rather than headline outcomes.
• Methods: instruments get redesigned for the population’s reality (language, literacy, trust), usually improving data quality for everyone.
• Collection: trained peer researchers reach respondents and candour professional outsiders cannot — at the cost of new confidentiality management.
• Analysis: joint sense-making sessions test interpretations against lived knowledge; findings that survive are sturdier, and the dissents are data.
• Use: recommendations co-authored by the people who must implement them skip the adoption chasm entirely — the practical stream’s core payoff.

A city housing NGO commissioned a participatory evaluation of its tenancy-support service, practical stream, with depth negotiated up front: six tenant co-evaluators (recruited openly, paid at researcher day-rates) holding joint control over questions, instruments, and interpretation; the NGO holding final report sign-off — stated in the contract everyone signed.

The tenants reshaped the evaluation immediately: to the commissioned questions about tenancy sustainment they added “what does the service cost tenants?” — time, disclosure, the feeling of being case-managed — a burden dimension the programme theory had never named. They redesigned the interview guide into plain speech, moved interviews out of the service’s offices, and conducted half themselves after a three-day training; disclosure rates about rent arrears, the topic tenants hide from caseworkers, were visibly higher in peer-led interviews.

Joint sense-making produced the evaluation’s sharpest finding and its recorded dissent: evidence that intensive check-ins sustained tenancies while making tenants feel surveilled split the panel — staff read it as a tolerable trade-off, tenant co-evaluators as a fixable design flaw. The report carried both readings, attributed; the board, confronted with the disagreement rather than a smoothed average, redesigned check-ins around tenant-controlled scheduling. Use, in the end, was the point — and the use followed the ownership.

• Participation as recruitment. Calling respondent interviews “participatory” — involvement means decision rights, not presence.
• Stream confusion. Promising transformation while running a utilisation design (or the reverse); the Cousins–Whitmore dimensions exist to prevent exactly this.
• Unpaid co-labour. Extracting evaluation work from participants for free reproduces, inside the evaluation, the power relation the method critiques.
• Capacity theatre. A two-hour “training” before sending peers to interview on sensitive topics — bad ethics and bad data in one move.
• Consensus laundering. Erasing interpretive disagreements to deliver a clean report; the disagreements were findings.
• Late participation. Inviting stakeholders to comment on a drafted report and branding the exercise participatory — the most common failure because it is the cheapest.

Participation costs time and money — training, sessions, compensation — and compresses badly: a genuinely participatory evaluation on a conventional timeline delivers conventional evaluation with extra meetings.

Co-evaluators are positioned actors: staff protect programmes, participants have stakes and factions. The design manages this with mixed panels, transparent selection, and recorded dissent — managed, never eliminated, and the report should say so.

And shared authority is only as real as the commissioning politics allow. Where funders hold every material decision, honest practice is to name the involvement consultation and deliver it well, rather than inflate the label — the field’s credibility problem is mislabelled participation, not modest participation.

Community-based participatory research extends shared control from one evaluation to a durable research partnership.

Most significant change is a story-based technique often embedded within participatory designs.

Photovoice gives participants the camera — a specific participatory data-and-advocacy method.

Utilization-focused evaluation shares the use rationale while keeping technical control with the evaluator.

Participatory processes multiply qualitative material — peer-led interviews, workshop recordings, disagreement logs — and the sense-making sessions need that material organised fast, in plain view. Evidano transcribes the interviews, codes them against the co-developed framework with every theme linked to quotable evidence, and gives sense-making workshops a shared, checkable evidence base instead of a facilitator’s summary; tenant co-evaluators can click from a theme to the words behind it, which is participation extended into the analysis itself.

The negotiation of power — who decides, whose reading prevails, what the report owns — is the method’s substance and belongs to the people in the room.