What is a logic model?
A logic model is a visual description of how a program’s resources and activities are expected to connect to its results. It helps a team agree what it will do, what it will deliver and what should change. A model describes the intended logic; it does not demonstrate that the results have occurred.
You may see four, five or six columns. Some formats separate short-, medium- and long-term outcomes; others use a final impact column. Follow the format you need while keeping the distinctions clear.
The components of a logic model
| Component | Question | Example |
|---|---|---|
| Inputs | What resources are available? | Trainers, equipment, partner time and funding. |
| Activities | What will the team do? | Deliver assessed practice and employer introductions. |
| Outputs | What will the work deliver? | Sessions completed and interviews arranged. |
| Outcomes | What changes for participants? | Stronger skills and suitable employment. |
| Longer-term aim or impact | What wider change does the work support? | More stable livelihoods. |
Context and assumptions belong around the model, even when they are not separate columns. A useful model explains the conditions needed for the arrows to make sense. CDC’s example shows a format with outcomes separated by time horizon.
A completed workforce logic model
This is an illustrative example, not a customer result. Adults with relevant experience need a recognized qualification and a route into suitable work.
| Resources | Activities | Outputs | Near-term outcomes | Later outcomes |
|---|---|---|---|---|
| Instructors; assessment tools; employer partners; transport support. | Provide practice, assess skills, help with applications and arrange interviews. | Practice sessions delivered; assessments completed; interviews held. | Participants demonstrate required skills and can explain them to employers. | Participants enter and sustain work that meets agreed suitability criteria. |
The pathway assumes that jobs are available, employers recognize the assessment, and participants can attend. If people learn the skills but cannot travel to available jobs, adding more instruction may not address the missing connection. The model should make that possibility visible.
How to build a logic model
- Define the problem and population. Avoid beginning with a favorite activity before understanding the need.
- Write the longer-term change, then the nearer-term outcomes that could support it.
- Choose activities that plausibly support those outcomes. Identify evidence or experience behind each connection.
- List the resources and outputs needed for delivery. Check whether the plan is feasible.
- Ask participants, delivery staff and partners what is missing. Include barriers and possible unintended effects.
- Keep the first diagram readable. Put detailed indicators and collection rules in a companion table.
Read the draft both forward and backward. Forward: if we do this, why would that follow? Backward: for this outcome to happen, what else must be true? Remove arrows that express hope without an explanation.
Add indicators without overloading the diagram
| Model element | Possible measure | Interpretation limit |
|---|---|---|
| Delivery | Number completing the assessment. | Completion is not demonstrated competence. |
| Learning | Share meeting a defined assessment rubric. | Report who was assessed and who was missing. |
| Employment | Share in suitable work at a named follow-up point. | Define suitability and disclose unknown status. |
| Experience | Accounts of barriers and what helped. | Do not treat a few interviews as a population percentage. |
Keep the baseline, target, source, timing and owner with each measure. For additional structure, use a results framework.
Common logic model mistakes
- Calling delivery an outcome: “100 people trained” describes an output; it does not establish a change in skill.
- Skipping intermediate change: training rarely leads directly to a broad social outcome without other steps.
- Ignoring external conditions: partners, policies and market conditions can affect results.
- Using arrows as proof: a plausible connection still needs investigation.
- Making the diagram too detailed: use supporting notes for definitions and evidence.
Use the model during implementation
At review time, compare delivery and outcomes. If attendance is low, investigate access and scheduling. If attendance is high but learning is weak, investigate the content, practice and assessment. If learning is strong but application is weak, examine the conditions for use.
Keep revisions dated and explain what changed. Documents, observations and interviews can help explain the pattern; retain their source and context. AI may help organize that material, but the interpretation still needs a reviewer.
Logic model and theory of change
A logic model commonly gives a compact view of resources, work and results. A theory of change makes the reasoning, assumptions and context more explicit. In practice the terms overlap. Use the required format and make sure the explanation is not lost. Continue to the theory of change guide.
Frequently asked questions
Is there one correct logic model template?
No. Formats vary. The model should make the work, expected results and important assumptions understandable.
Should impact be a separate column?
It can be, but some formats use long-term outcomes instead. Define the terminology rather than adding a column solely to reach a fixed number.
Can a logic model include negative effects?
Yes. Include possible unintended or harmful effects in the assumptions, risks or outcome discussion and decide how to monitor them.
Does every activity need a separate outcome?
Not necessarily. Several activities may support one outcome, and one activity may support several outcomes. Keep the relationships plausible and readable.
