Choosing between a baseline, needs assessment and feasibility study

Organizations often request a baseline, needs assessment, or feasibility study as if the three were interchangeable. They are not. Each starts from a different uncertainty, supports a different decision, and produces a different kind of evidence.

Choosing the wrong study can create a technically polished answer to the wrong question. A baseline may describe conditions without establishing which response is most appropriate. A needs assessment may identify serious gaps without showing whether a proposed intervention can operate. A feasibility study may test a proposed model without providing the measures needed for later comparison.

The simplest way to choose is to begin with the decision, not the study label.

The three questions in plain language

A baseline asks: What is the starting condition against which future change will be examined?

A needs assessment asks: What needs, gaps, risks, capacities, and priorities exist now, and for whom?

A feasibility study asks: Can a proposed intervention work in this context, under these conditions, with acceptable risks and resources?

These questions can overlap. The difference lies in the primary use of the evidence. A commissioner should be able to complete this sentence before writing terms of reference: “When this study is finished, we need to decide whether or how to...”

If that sentence is unclear, the study is not ready to be designed.

When a baseline is the right choice

A baseline establishes a credible starting point for later comparison. It is normally connected to a programme's results framework, theory of change, indicators, target population, and planned measurement schedule.

A useful baseline does more than record values at the beginning. It tests whether indicators are defined clearly, whether required data can be collected responsibly, whether the sample supports the comparisons the programme expects to make, and whether assumptions in the programme design remain plausible.

A baseline is usually appropriate when:

• The intervention and intended population are sufficiently defined
• The programme has indicators or learning questions that require a starting measure
• Future monitoring, midline, endline, or evaluation will use comparable definitions
• The commissioner needs to understand variation across meaningful groups or locations
• Data collection can occur before substantial programme exposure changes the starting condition

A baseline is not automatically the first research activity a new programme needs. If the problem, priority population, or intervention model is still uncertain, a needs assessment or feasibility study may need to come first.

Common baseline mistakes

The first mistake is measuring everything in a results framework without testing whether each indicator is useful for a decision. The second is designing the baseline after implementation has already influenced the intended participants. The third is collecting a large amount of contextual data without a clear plan for later comparison. The fourth is treating an initial value as evidence that the programme caused future change.

A baseline supplies a reference point. On its own, it does not establish attribution.

When a needs assessment is the right choice

A needs assessment builds a structured understanding of current conditions. It may examine the scale and severity of needs, who is most affected, how needs vary, what services or capacities already exist, which barriers matter, and how people understand their own priorities.

Needs assessments can support humanitarian prioritization, programme design, targeting, service adaptation, resource allocation, or further inquiry. They can be rapid or extensive. The appropriate depth depends on the urgency, existing information, risks, access, population diversity, and consequence of the decision.

A needs assessment is usually appropriate when:

• The current situation is insufficiently understood
• The commissioner needs to compare needs across groups, locations, or sectors
• Existing evidence is fragmented, outdated, or not suitable for the decision
• Programme priorities or targeting criteria need to be established or reviewed
• The study must examine both gaps and existing capacities
• An evolving context requires an updated picture before action

The UNHCR guidance on data and information management emphasizes that operational information should help identify needs, gaps, risks, and response priorities. That is a useful discipline for a needs assessment: collect information because it supports an operational choice, not because it might be interesting later.

Common needs-assessment mistakes

One mistake is equating the most visible or frequently reported problem with the highest priority. Another is collecting only deficit information while ignoring local capacities, current coping strategies, existing services, and people's preferences. A third is combining incomparable sources without explaining their timing, definitions, or limitations. A fourth is assuming that expressed need alone determines programme feasibility.

A needs assessment should not promise precision that access, sampling, timing, or measurement cannot support. It should clearly distinguish direct observations, participant reports, administrative records, estimates, and interpretation.

When a feasibility study is the right choice

A feasibility study tests whether a proposed intervention is workable. It examines the conditions required for implementation and the barriers that may prevent the model from operating safely, acceptably, and effectively.

Feasibility is wider than logistical possibility. A programme might be physically deliverable but unacceptable to intended participants. It might be acceptable but financially unsustainable. It might be affordable but dependent on a market, workforce, supply route, technology, permission, or service capacity that is not reliable enough.

A feasibility study is usually appropriate when:

• A specific intervention or delivery model is under consideration
• Decision makers need to compare implementation options
• Important assumptions about access, capacity, markets, technology, cost, or acceptance remain untested
• The organization needs to identify operational, ethical, environmental, or conflict-related risks
• A pilot, phased approach, adaptation, or decision not to proceed are credible outcomes
• The programme needs practical conditions and safeguards before investment

Common feasibility-study mistakes

The most serious mistake is commissioning the study only to justify a decision already made. This narrows inquiry and discourages uncomfortable findings. Another is examining logistics while ignoring participant acceptance, power, protection, data responsibility, environmental implications, or unintended effects. A third is presenting a single recommendation without showing the conditions, tradeoffs, and uncertainty behind it.

A credible feasibility study permits several conclusions: proceed, proceed with changes, pilot and reassess, delay until conditions change, or do not proceed.

Use evaluation thinking before the programme begins

The OECD guidance on applying evaluation criteria thoughtfully cautions against treating relevance, coherence, effectiveness, efficiency, impact, and sustainability as a mechanical checklist. Criteria should be selected and interpreted according to the purpose, users, context, and stage of the work.

That principle is useful even before a formal evaluation.

A needs assessment often concentrates on relevance: which needs matter, to whom, and whether the proposed direction responds to them. It may also consider coherence by examining existing services, policies, actors, and possible duplication.

A feasibility study may test likely effectiveness, operational efficiency, coherence, and sustainability before implementation. It asks whether the conditions required for the proposed results are present or can realistically be created.

A baseline establishes measures that may later contribute to judging effectiveness or impact. Its design should anticipate future interpretation without claiming that a before-and-after difference will explain causation by itself.

The criteria help sharpen questions. They do not determine the study type automatically.

When one study needs elements of another

Real commissioning decisions rarely fit into neat boxes. A programme may need a needs assessment with a feasibility component. A feasibility study may include a limited baseline for indicators that must be measured before implementation. A baseline may reveal design assumptions that require additional qualitative inquiry.

Combining elements is reasonable when each component has a clear purpose. It becomes risky when several study labels are merged into one assignment without adjusting time, expertise, sample, fieldwork, analysis, or budget.

For example, a request for “a baseline and needs assessment” should specify:

• Which decisions belong to programme design and which belong to future measurement
• Whether the same population and sample can answer both sets of questions
• Which indicators require statistical estimates
• Which needs require qualitative or participatory understanding
• How findings from each component will be reported and used

Without this separation, the assessment can become broad but shallow.

A practical selection test

Before choosing a study, answer six questions.

1. What decision follows?

Name the actor, choice, and timing. “Inform programming” is too broad. “Decide which two service models to pilot before the next funding cycle” is more useful.

2. What is already known?

Review existing studies, monitoring records, administrative data, contextual analysis, and lived knowledge. New primary data should fill an important gap, not repeat what is already sufficient.

3. What remains uncertain?

Separate uncertainty about current needs, the proposed model, and the starting measure. This usually reveals the primary study type.

4. What level of confidence is needed?

A rapid operational choice and a major long-term investment may require different depth, sample, review, and assurance. Match the design to the consequence of being wrong.

5. Who could be affected by the study or decision?

Consider participation, burden, consent, exclusion, protection, data sensitivity, and how findings could be used or misused. The UNICEF approach to ethical evidence generation reinforces that ethics belongs throughout evidence work, not only at an approval stage.

6. How will the answer be used?

Decide whether users need a dataset, planning workshop, targeting framework, implementation conditions, indicator reference, concise brief, or formal report. The output should fit the decision process.

Final choice

Choose a baseline when the central need is a credible starting measure for future comparison.

Choose a needs assessment when the central need is to understand and prioritize current needs, gaps, risks, and capacities.

Choose a feasibility study when the central need is to test whether and under what conditions a proposed intervention can work.

If all three seem necessary, separate the decisions first. Then design only the components that produce evidence those decisions will use.

Planning checklist

• The decision and intended users are named
• The main uncertainty is explicit
• Existing evidence has been reviewed
• The study type matches the primary question
• Secondary components have distinct purposes
• Methods are proportionate to the decision
• Ethical and data risks are considered
• Limitations and inference boundaries will be reported
• Outputs fit the real decision process
• Time and resources match the requested scope

Further reading

• OECD: Applying Evaluation Criteria Thoughtfully
• UNHCR: Data and Information Management
• UNICEF Innocenti: Ethical Evidence Generation

Closing call to action

Heading: Not sure which study your decision needs?

Body: ERC can help clarify the decision, evidence gap, and proportionate study design before methods are fixed.

Action: Discuss a study