Mapping Stakeholder Evidence for a Policy Study
Map whose evidence a policy study needs, including affected groups and missing perspectives, without treating access or influence as proof.
Stakeholder evidence mapping is the disciplined connection of each part of a policy question to the people, records and organisations that can answer it, together with an explicit account of how that evidence will be collected, weighed and reported, and of what it will leave unresolved. The output is a working map, not a consultation roster: it should show which source speaks to which claim, what that source can bear, and where the study will record uncertainty rather than fill it.
What the map has to settle
A stakeholder map is useful when it settles four design questions. First, whose experience bears on the specific effect the option is meant to produce, and whose experience is adjacent but not probative. Second, what each source can support: a claim about how a process works, a claim about how often a problem occurs, or a claim about magnitude. Third, how disagreement between accounts will be recorded and weighed. Fourth, what the study will not answer, stated before fieldwork rather than after it.
Treat these as design commitments. A list of groups without a corresponding claim leaves the fieldwork without a test of relevance.
Start from the decision, not the organisation chart
The policy question and the evidence question are different. "Should the standard apply to small providers?" is a policy question. "What would we need to know about compliance cost and service quality to judge that?" is an evidence question. The map should be built from the second.
The UK Government Policy Profession's standards annex, a capability framework for the UK Civil Service rather than a requirement in other jurisdictions or for external researchers, treats problem definition, working with multiple kinds of evidence, research methods, stakeholder perspectives and the assessment of policy outcomes as connected skills (Policy profession standards: annex). The useful implication for mapping is structural: stakeholder perspectives belong to the same chain as the problem definition and the outcome test, not to a separate consultation phase.
A scoping sheet as a working format
The following seven fields are a suggested working format. They are not a validated instrument and not an institutional requirement; adapt the labels to the study. Their purpose is to make each row of the map carry a claim, a source and a limit.
- Problem. State the observed or alleged condition the option addresses, and separate it from the option itself. If the problem is contested, say who contests it and on what basis.
- Option, including the baseline. Describe the proposed change and the counterfactual: what happens now, or what would happen without the option. Without a baseline, stakeholder accounts of burden cannot be interpreted.
- Affected group. Name groups by their relationship to the effect, not by generic labels. "Providers required to submit the return" and "staff who complete it" may hold different evidence.
- Effect pathway. Set out the steps between the option and the outcome in sequence: what changes, for whom, through what mechanism. Each step is a candidate for evidence.
- Evidence. For each step, name the source and the access route: interviews, administrative records, inspection data, published research, a timed exercise.
- Uncertainty. Name the type of uncertainty: unknown magnitude, contested mechanism, missing baseline, unverified claim. "We do not know" is more useful when it says what is unknown.
- Follow-up. Record how the gap will be closed, by whom and when, or state that it will remain open.
Match the claim to what the source can bear
Different sources support different claims. Experiential accounts are strong on mechanism — how a requirement is actually completed, where discretion is exercised, which steps absorb time. They are weak on prevalence unless the sample is designed for it. Administrative records are strong on counts and timing where definitions are stable, and weak on why a figure changed. Professional judgement is useful for feasibility and for identifying effects that are hard to measure, and should be attributed as judgement rather than presented as measurement.
An account of mechanism cannot answer a question of magnitude. A provider describing a difficult return does not establish how many providers find it difficult. When the decision turns on magnitude, the map needs a source that can bear that claim, or an explicit statement that the study will not estimate it.
Select for contrast or for representation, and say which
Sampling follows from the claim. If the decision turns on how an effect varies across different conditions (size, location, delivery model), select contrasting cases and report variation. If it turns on how common an effect is, the study needs a sampling frame that supports inference, or it must state that prevalence is outside its scope. These are conditions rather than preferences: the right choice depends on the decision the study is meant to inform.
Access and ethics constrain the map as much as method does. Consent should cover how accounts will be used and whether they will be attributed. Where commercial sensitivity or fear of retaliation is plausible, anonymity needs to be designed in before the first conversation rather than promised afterwards. Collect only what the claim requires. Mapping should not become a route to gathering private client material or identifying details the work does not require.
Weighting, disagreement and asymmetry
Weigh accounts by proximity to the effect pathway, corroboration and incentive. Proximity: the person who completes a form speaks with more authority about its operational burden than a representative body commenting on it. Corroboration: an account that matches records or a second independent source carries more weight than a single claim. Incentive: a regulated party and a regulator may both describe the same process accurately while emphasising different costs.
Do not average disagreement into a misleading middle. Record the competing accounts, state what would distinguish them, and where possible add a follow-up that could. Asymmetry deserves the same treatment. Frontline staff may have less standing to contradict an employer's account. Small organisations may have less capacity to participate at all. The map should note whose evidence is likely to be missing and what that omission does to the conclusions.
A fictional worked example
The following proposal is fictional and is used only to show the sheet in use. It is not a real standard, programme or finding.
Suppose a fictional department proposes that small residential support providers submit a quarterly staffing return, replacing an annual return, to improve workforce planning. The policy team estimates that the new return will take about thirty minutes to complete. The map does not accept that estimate; it records it as unverified.
| Affected group | What this group can evidence | Limits |
|---|---|---|
| Provider managers | How the return is compiled, who assembles it, what systems hold the data | May generalise from their own organisation; commercial sensitivity about staffing |
| Frontline staff | The point at which data collection competes with direct support | Limited view of reporting requirements as a whole |
| Local authority contract officers | How returns are used, what triggers follow-up, what data quality problems recur | Indirect experience of provider burden |
| Payroll or rostering software suppliers | Whether standard reports can produce the required fields | Interested in the design outcome |
For the effect pathway, the fictional sheet records: quarterly frequency increases the number of reporting cycles; the new fields require data not held in one system; compilation time falls on a manager or administrator; time displaces other duties. The baseline is the existing annual return, whose completion time is not currently measured.
Uncertainty is recorded in three places: the magnitude of compilation time, whether the required fields can be produced from existing systems, and whether the displacement effect is material for small providers. Follow-up is a timed exercise with a small number of volunteer providers using the draft return and a think-aloud protocol, plus a request for the existing annual return's completion records. If the department cannot obtain a representative sample, the sheet states that the study will describe mechanisms and variation but will not estimate average burden. That sentence is more useful to a decision-maker than a number with no basis.
Checks before fieldwork or commissioning
Five checks are worth applying. Every group affected by the option has a row, or a stated reason for exclusion. Every evidence entry names a source and an access route rather than a topic. The effect pathway contains at least one step that could be observed or measured. Each uncertainty is specific enough to guide a follow-up. The brief states what the study will not answer.
These checks also apply when the study is commissioned. A brief that specifies only the topic and the number of interviews leaves the mapping to the supplier and makes the output hard to assess. A brief that includes the decision, the option and baseline, the effect pathway and the intended use of each source gives a research partner something to design against. Policy Expert Network's guide to assessing an expert project before accepting it covers the reciprocal questions a researcher should ask about scope and access.
Boundaries
The sheet is a scoping and communication device. It does not replace formal impact assessment, legal analysis, statistical design or ethics review, and it does not establish prevalence on its own. Its value lies in making the evidential basis of a policy study visible early enough to change the design.
To begin, take one live option and complete the seven fields for the group closest to the effect, then work outward. A field that cannot be completed is not a gap in the worksheet; it is the first research task.