Theory of Change: Meaning, Model and Education Examples

A Theory of Change explains how and why a specific intervention is expected to contribute to a series of outcomes. It identifies the proposed links between action and change. It also records the assumptions, evidence and outside conditions that may support or disrupt each link (Weiss, 1995).

The word theory matters. A pathway sets out ideas to test, not proven effects. A diagram can make causal reasoning visible, but it cannot show that an intervention caused an outcome.

That conclusion needs a suitable evaluation design and evidence that tests other explanations (HM Treasury and Evaluation Task Force, 2026).

This article uses Theory of Change for the explanation and logic model for its structured form. Other authors and organisations use the terms differently. Their overlap should be clear.

For other accounts of learning and change, see the Learning Theories collection. Theory of Change belongs chiefly to programme and policy evaluation. It is not a psychological theory of learning.

‌

Direct Definition and Evidence Boundary

A Theory of Change describes how an intervention may contribute to intended or observed results. The intervention can be a policy, programme, service or school-improvement project. The theory applies to that intervention and the setting in which it operates.

At minimum, it should make the proposed sequence of change clear. It may begin with resources and activities. It can then pass through outputs and near-term outcomes before reaching long-term outcomes or impacts. It should also explain why each link may hold and under what conditions.

The OECD definition asks how people understand change in a given setting. This includes their assumptions about links between inputs, activities and results. The UK government's Magenta Book also covers evidence strength, wider context, causal mechanisms, other routes and risks (OECD, 2023; HM Treasury and Evaluation Task Force, 2026).

This makes Theory of Change useful for inquiry. A team can ask whether the intervention was delivered and whether the expected changes occurred. It can test key assumptions and consider other causes of the result. The model does not supply answers in advance.

‌

Components and Construction Without a Universal Template Claim

There is no universal six-part template. A short Theory of Change may use a brief account and a small diagram. A policy evaluation may need several linked models, an evidence review and a list of assumptions. The right level of detail depends on the intervention, its setting and the evaluation questions.

A defensible account usually addresses the following elements:

1. Problem and context. What problem should the intervention address? Who is affected, and under what social, organisational or policy conditions?
2. Intended results. What outputs, near-term outcomes and longer-term outcomes are expected? Unintended and negative outcomes should also be considered.
3. Activities and resources. What will the intervention provide or do, and what capacity is required?
4. Causal links or mechanisms. Why are particular activities expected to produce particular changes?
5. Assumptions and external influences. What conditions must hold? What else could enable, prevent or independently produce the outcome?
6. Evidence and uncertainty. What supports each link, where is the evidence weak and which rival explanations remain plausible?
7. Indicators and evaluation questions. What would show whether delivery and near-term outcomes occurred? What methods are needed to assess contribution or effect?

A team can use this short sequence without treating it as a fixed template:

1. Define the problem and setting. State who is affected and what lies inside or outside the intervention.
2. Set out the intended results. Link activities and outputs to near-term and long-term outcomes.
3. Explain each proposed link. Record the mechanism, assumptions, outside influences and current evidence.
4. Choose what to test. Identify the most important uncertainties, rival explanations, indicators and evaluation questions.
5. Review the model. Revise or reject links when delivery data, outcome evidence or context challenges them.

Assumptions are more than hopeful background statements. They may concern a proposed causal link or the conditions needed for it to operate. Mayne argues that evaluators should state each assumption. They should then select evidence that fits the specific causal question (Mayne, 2023).

Indicators also need a narrow interpretation. An attendance measure may show that participants attended training. A practice measure may show that a new routine appeared. Neither fact alone demonstrates that training caused the practice change or that the practice change caused a later outcome.

Teams can start with a desired outcome and work back to possible preconditions. They can also start with an existing programme and state its current logic. Backward mapping is a common method, not the only correct one.

A Theory of Change may be built before, during or after delivery. It should change when evidence contradicts it.

‌

Theory of Change Study Notes

Read the study note as text

Core idea. A Theory of Change explains how and why a specific intervention is expected to contribute to outcomes in a particular context. The pathway is a hypothesis to test, not proof of impact.

Model anatomy. Activities connect to near-term outcomes, later outcomes and longer impact. Every arrow needs assumptions, context, evidence and rival explanations.

What to specify. State the problem, people and context; intended and unintended results; activities, outputs and resources; proposed mechanisms and links; assumptions and external factors; evidence, uncertainty and indicators.

Related terms. A logic model is a structured programme map and may include assumptions and context. Programme theory is a broad explanation of how a programme should work. A results chain links activities to outputs, outcomes and impacts.

Education use. Use Theory of Change for a defined programme, policy, inspection framework, professional-development initiative or school-improvement strategy. Ordinary lesson planning does not require one.

Limitations. A plausible diagram can be false; agreement can hide power differences or false consensus. Indicators do not establish attribution by themselves, and complexity, weak evidence or omitted context can undermine a model.

Memory cue. Explain, test, revise.

‌

Theory of Change, Logic Model, Programme Theory and Results Chain Comparison

Writers often offer a neat distinction: logic models show what a programme does, while Theory of Change explains why it should work. That can be a useful working convention, but it is not a settled disciplinary rule. Reinholz and Andrews note that logic models may also include assumptions and contextual factors, making their distinction from Theory of Change unclear in practice (Reinholz and Andrews, 2020).

Swipe horizontally to read the full table.

The OECD describes a results chain as the causal sequence through which an intervention moves from inputs and activities to outputs, outcomes and impacts. A results framework commonly adds indicators, assumptions and risks. A Theory of Change can contain such a chain while also explaining the reasoning and evidence behind it (OECD, 2023).

These tools are therefore better understood as overlapping representations and practices than as a ladder from weak to strong. A concise logic model may communicate a well-evidenced programme clearly. A visually complex Theory of Change may still contain unsupported arrows.

‌

Origins and Disciplinary Context

The modern use of theories of change in evaluation grew from programme theory and theory-based evaluation. In a key 1995 chapter, Carol Weiss focused on large community projects and the theories beneath them. She argued that programme assumptions often stayed hidden and escaped close study (Weiss, 1995).

Weiss did not claim to have created every form of theory-based evaluation. Nor did she present Theory of Change as a school teaching method. Her chapter addressed complex community projects for children and families. It asked how evaluation could trace the small steps and assumptions through which programmes were expected to work.

Connell and Kubisch later developed the approach for large community projects linked to the Aspen Roundtable (Connell and Kubisch, 1998). Later work extended its use across public policy, global development, health, research and education.

The history is broader than a story in which Theory of Change replaced a weak logic model. Programme theory, theory-driven evaluation, programme logic and intervention logic have shared roots. Patricia Rogers described programme theory as a family of ways to link programme actions to outcomes. The models can then inform evaluation (Rogers, 2008).

‌

Theory of Change Versus Change Theory and Change Management

A Theory of Change for one project is not a general theory about change. Reinholz and Andrews define change theory as research-backed ideas that explain change beyond a single project. A project-specific model may draw on several such theories. They can help explain its setting, mechanisms and prior conditions (Reinholz and Andrews, 2020).

For example, research on how organisations learn may show when staff development becomes part of routine work. That research can inform an assumption in the project's Theory of Change. It does not set the whole pathway. The project still has its own people, actions, setting and goals.

Change management is also distinct. It covers the practices used to prepare, lead and sustain change in an organisation. Leadership routines, clear messages, staff input and delivery plans may help a team act.

Those practices may inform the model's assumptions. They do not explain why the intervention should produce its intended outcomes.

Readers seeking the broader organisational traditions should use the separate guide to change theories. A systems theory perspective may also inform a complex intervention, but Theory of Change is not itself one universal systems theory.

‌

One Bounded Education Programme Example

Consider a fictional trust-wide programme to improve feedback on extended writing. It provides subject-based staff development, joint planning time and short coaching cycles. Its long-term goal is stronger learner writing. The pathway below is a claim to test, not evidence that the programme works.

Swipe horizontally to read the full table.

Several assumptions could fail. Staff may lack protected time, coaching may vary in quality, the agreed approach may not suit every subject, learners may not understand the feedback or assessment changes may affect writing scores. These conditions should be recorded and examined rather than hidden behind arrows.

The distinction between implementation and effect is essential. If the professional development was not delivered as intended, the programme has not received a fair test. If delivery and practice changed but learner writing did not, the causal reasoning or outcome measures may need revision. If writing improved, evaluators must still consider other plausible causes.

EEF guidance asks teams to state their causal and contextual assumptions. Where possible, they should support each link with existing evidence or education theory. The evaluation can then test selected links (Education Endowment Foundation, 2022).

Ofsted's plan for the Education Inspection Framework gives a policy example. It separates inputs, mechanisms, short-term outcomes and unintended effects (Ofsted, 2023).

This use fits a wider commitment to evidence-informed practice. A school does not need a new causal map for every lesson. Leaders may use the framework for a defined and major project. This may warrant shared assumptions, monitoring and evaluation within a whole-school policy process.

‌

Uses in Planning, Monitoring, Evaluation and Learning

During planning, a Theory of Change can expose missing steps, weak evidence and incompatible expectations before implementation begins. It can also show where the intervention depends on another service, policy or group. This is stress-testing, not a guarantee of success.

During monitoring, the model can help distinguish delivery information from outcome information. It gives teams a reasoned basis for choosing indicators at different points in the pathway. Monitoring may reveal that an assumption is not holding, but it does not by itself establish the intervention's causal effect.

During evaluation, the theory can guide questions about delivery, mechanisms, context, outcomes and unintended effects. Different links may need different methods. Process evaluation can examine delivery and experience, while impact evaluation can assess effects.

Contribution analysis or process tracing may suit some causal questions when trials are not feasible. Their limits must remain clear (HM Treasury and Evaluation Task Force, 2026).

During learning, findings can be used to revise the model. An intervention may work for one group but not another, or only when a contextual condition is present. A useful Theory of Change should be capable of being wrong and capable of changing when evidence accumulates.

Evidence about the use of the approach remains cautious. Breuer and colleagues identified 62 public-health papers using Theory of Change, but reporting of its development and use often lacked detail. Their review demonstrates uptake and reporting problems; it does not establish that Theory of Change makes interventions more effective (Breuer et al., 2016).

‌

Limitations and Common Failure Modes

A plausible model can still be false. A team may agree that one step leads to another without enough evidence. Shared agreement can help make a claim clear, but it does not validate that claim. Each proposed link must remain open to evidence and other explanations.

Arrows can hide uncertainty. A line between an activity and an outcome may hide several mechanisms, local conditions and rival explanations. Mayne's work on assumptions is useful because the content of each causal link needs close study (Mayne, 2017, 2023).

Participation can create false consensus. Stakeholders may hold different views, interests and levels of influence. Weiss foresaw difficulty and political risk when programmes made assumptions clear.

Group work can expose differences. It does not remove power or guarantee that a shared model is right (Weiss, 1995).

Complex diagrams can imitate rigour. More boxes and arrows do not necessarily mean better reasoning. Complex interventions may need feedback loops, multiple pathways and system boundaries, but each addition should serve an evaluative purpose.

Models can neglect context. Lam's review finds several recurring concerns. They include vague concepts, weak models, poor attention to context and complexity, and uncertain stakeholder input (Lam, 2020).

The theory may never be tested. A model used only in a funding proposal becomes decorative. Indicators, data and evaluation methods must be connected to the most important uncertainties and assumptions.

The model may be protected from revision. If every outcome is explained as success and every failure as poor implementation, the theory cannot be challenged. Evaluators should consider negative programme theory, rival explanations and evidence that would count against the proposed pathway.

School improvement claims need care. Coe notes that cause and transfer remain problems in school-improvement evidence. A Theory of Change can clarify a claim. It cannot correct weak measures, selective reports or the lack of a sound comparison (Coe, 2009).

‌

‌

Frequently Asked Questions

These concise answers preserve the article's central boundary: a Theory of Change is an intervention-specific explanation that can organise inquiry, but the model does not establish causal effect without suitable evaluation evidence.

‌

What is a Theory of Change?

In brief, it is the working explanation behind a defined intervention. Its links, assumptions and conditions remain claims to test.

‌

Does a Theory of Change prove that a programme works?

No. It expresses an expected pathway. Evidence from an appropriate evaluation is needed to determine whether activities occurred, outcomes changed and the intervention contributed to those changes.

‌

How is a Theory of Change different from a logic model?

One common convention is that a logic model summarises resources, activities, outputs and outcomes, while a Theory of Change gives greater attention to why links should hold. In practice, logic models may also contain assumptions and causal explanations, so the terms overlap.

‌

What is the difference between Theory of Change and change theory?

A Theory of Change is normally specific to one initiative and context. Change theory is more general knowledge about how change occurs across cases. Relevant change theory can strengthen the assumptions in a project-specific model.

‌

Is Theory of Change a change-management model?

No. Change management concerns how organisational change is led and implemented. Its research and practices may inform part of a Theory of Change, but they do not replace the intervention's causal explanation or evaluation.

‌

How can schools use Theory of Change?

A school or trust can use it to specify and evaluate a defined programme or improvement initiative. The team should distinguish expected links from evidence, examine assumptions and plan suitable measures. Ordinary lesson planning does not require a Theory of Change.

‌

When should a Theory of Change be revised?

It should be reviewed when implementation data, outcome evidence, stakeholder experience or contextual change challenges an assumption. Revision is a sign that the model is being used for learning rather than treated as a fixed promise.

‌

Further Reading and References

These primary, peer-reviewed and authoritative sources support the article's history, methods, terminology and education boundaries. They also establish the central caution that a Theory of Change organises claims and evaluation questions but does not prove programme impact by itself.

• Weiss, C. H. (1995). Nothing as Practical as Good Theory: Exploring Theory-Based Evaluation for Comprehensive Community Initiatives for Children and Families. In New Approaches to Evaluating Community Initiatives. Aspen Institute.
• Connell, J. P., & Kubisch, A. C. (1998). Applying a Theory of Change Approach to the Evaluation of Comprehensive Community Initiatives: Progress, Prospects, and Problems. In New Approaches to Evaluating Community Initiatives: Theory, Measurement, and Analysis, Volume 2. Aspen Institute.
• Rogers, P. J. (2008). Using Programme Theory to Evaluate Complicated and Complex Aspects of Interventions. Evaluation, 14(1), 29-48.
• Coryn, C. L. S., Noakes, L. A., Westine, C. D., & Schröter, D. C. (2011). A Systematic Review of Theory-Driven Evaluation Practice From 1990 to 2009. American Journal of Evaluation, 32(2), 199-226.
• Mayne, J. (2017). Theory of Change Analysis: Building Robust Theories of Change. Canadian Journal of Program Evaluation, 32(2), 155-173.
• Mayne, J. (2023). Assumptions in Theories of Change. Evaluation and Program Planning, 98, 102276.
• Reinholz, D. L., & Andrews, T. C. (2020). Change Theory and Theory of Change: What's the Difference Anyway? International Journal of STEM Education, 7, 2.
• Breuer, E., Lee, L., De Silva, M., & Lund, C. (2016). Using Theory of Change to Design and Evaluate Public Health Interventions: A Systematic Review. Implementation Science, 11, 63.
• Lam, S. (2020). Toward Learning from Change Pathways: Reviewing Theory of Change and Its Discontents. Canadian Journal of Program Evaluation.
• Coe, R. (2009). School Improvement: Reality and Illusion. British Journal of Educational Studies, 57(4), 363-379.
• OECD. (2023). Glossary of Key Terms in Evaluation and Results-Based Management for Sustainable Development. Second edition.
• HM Treasury and Evaluation Task Force. (2026). Magenta Book: Central Government Guidance on Evaluation.
• Education Endowment Foundation. (2022). Guidance on Theory of Change and Logic Models.
• Ofsted. (2023). EIF Evaluation Plan and Theory of Change.