Guide
How to build a logic model for a grant
A step by step guide to building a logic model for a grant application or evaluation plan, with an example from a youth prevention program and common mistakes.
In short
A logic model is a one page picture of how your program is supposed to work: the resources you have, the activities you will run, what those activities produce, and the changes you expect as a result. Build it from right to left: start with the outcome you want, then work back to the activities and resources that will get you there.
Key takeaways
- Start with the long term outcome and work backward.
- Keep outputs (what you do) and outcomes (what changes) separate.
- Every outcome needs at least one indicator you can measure.
- One page is the goal. If it does not fit, the program is not clear yet.
What goes in a logic model
Most logic models use five columns:
| Column | What it means | Example from a youth vaping prevention program |
|---|---|---|
| Inputs | What you have to work with | Coalition members, grant funds, school partners |
| Activities | What you will do | Classroom sessions, a parent campaign, retailer education |
| Outputs | What the activities produce, counted | Number of sessions held, students reached, retailers visited |
| Short term outcomes | Early changes | More students see vaping as risky |
| Long term outcomes | The change you are working toward | Fewer students report vaping in the past 30 days |
Step 1: Start with the outcome
Write down the long term change you want, as specifically as you can. "Reduce youth vaping" is a start. "Reduce past 30 day vaping among high school students in our town" is something you can measure.
Step 2: Work backward
Ask what has to change first for that outcome to happen. For vaping, research points to things like how risky young people think vaping is and how much they believe their parents disapprove. Those become your short term outcomes.
Step 3: Choose activities that move those outcomes
Each activity should connect to at least one short term outcome. If an activity does not connect to anything, either drop it or add the outcome it is meant to change.
Step 4: Count the outputs
Outputs are the countable products of your activities. They show you did the work. They are not proof that anything changed, which is why the outcomes columns matter.
Step 5: Add indicators
For every outcome, name how you will measure it. A youth survey can measure perceived risk and past 30 day use. Program records can measure outputs. If you cannot measure an outcome, rewrite it until you can.
Common mistakes
- Mixing outputs and outcomes. "Held 20 sessions" is an output. "Students who attended are more likely to see vaping as harmful" is an outcome.
- Promising too much. A one year grant rarely changes community wide rates. Put realistic short term outcomes in the first years.
- Building it alone. The people running the program should recognize their work in it.
Get help with yours
We build logic models and evaluation plans with grantees every year. If your application is due soon, get in touch.
Frequently asked questions
Is a logic model the same as a theory of change?
How detailed should a logic model be?
Do funders require a specific format?
Sources
- Logic Model Development Guide, W.K. Kellogg Foundation (2004) (opens in a new tab)
- Program evaluation framework, Centers for Disease Control and Prevention (opens in a new tab)