What is a theory of change? How to develop one and distinguish it from a logic model

There is no uniform boundary between the two. A logic model is one way to diagram a program's flow and can also represent a theory of change. When examining a theory of change, focus on the assumptions and supporting evidence behind each arrow.

Choose the necessary level of detail according to uncertainties in decisions, rather than program scale alone. You can share a concise diagram and add written explanations of important pathways.

Here we propose backward mapping: work back from long-term goals to the necessary conditions. Compare the pathway with actual existing practice and prior research so that it does not rest on aspirations alone.

Distinguishing what the program directly controls from what depends on the external context helps avoid excessive causal claims in evaluation.

1. Specify the long-term desired condition
2. Identify intermediate outcomes that must come first
3. Organize the sequence and relationships between intermediate outcomes
4. State the assumptions required for each change
5. Place the points where the intervention acts on the diagram
6. Turn hypotheses that need testing into evaluation questions

It is important not to try to demonstrate the entire theory of change at once. First select important hypotheses, high-risk hypotheses and hypotheses that stakeholders will use for decisions.

For example, if an important hypothesis is that ongoing support changes organizational decision-making, you need to examine changes in meeting structures, budget allocation and decision processes rather than training satisfaction.

• Turn hypotheses into evaluation questions
• Check whether assumptions still hold
• Distinguish outcome time horizons
• Combine quantitative data and qualitative information
• Update the theory during implementation

Before using a theory of change in evaluation, color-code what the program can directly do and what depends on the external context. This makes later explanations easier. Clarify the scope of the program's contribution to avoid overstating effects.

Not every change pathway needs to be measured with the same precision. Prioritize pathways directly relevant to decisions, those where failure would have major consequences, and those on which stakeholders disagree.

• The long-term goal describes a condition for the target group
• Assumptions are stated for each outcome
• Intervention points and external factors are distinguished
• Important hypotheses have been turned into evaluation questions