Build a logic model for your NSF, NIH, or nonprofit grant
A logic model shows reviewers how your inputs and activities lead to outputs and outcomes. Pick your funder and GrantCopilot gives you the columns it expects, drafts NSF and nonprofit models from what you have already written, and exports the model with your proposal.
What goes in a logic model
Starting point
What the model is built around. NSF starts from a goal and research questions, NIH from a program goal and Specific Aims, and nonprofits from the problem or need.
Inputs
The resources the project draws on: staff, faculty, volunteers, partners, facilities, funding, and prior work.
Activities
Five to seven high-level, action-oriented statements that start with a verb such as run, develop, or train. They are not a task list.
Outputs
What the activities produce. Things you count (people served, sessions held) and things you create (a curriculum, a manual, a program).
Outcomes
The changes that follow, in three steps: short-term during the project, mid-term near or just after it ends, and long-term after it.
Impact
NIH and nonprofit formats end with impact: the broader change for the field, the workforce, or your community. Do not promise to measure it.
Three formats, one for each kind of funder
NSF Theory of Change
The goal-to-outcomes chart NSF reviewers expect, built around your research questions.
Starts from
Your goal and up to 3 research questions
Columns
- Inputs
- Activities
- Outputs
- Short-term outcomes
- Mid-term outcomes
- Long-term outcomes
GrantCopilot drafts the columns from your goal, questions, and linked proposal.
NIH program logic model
For R25, training, and center grants. Each activity ties back to a Specific Aim.
Starts from
Your program goal and Specific Aims
Columns
- Specific Aims
- Inputs
- Activities
- Outputs
- Outcomes
- Impact
You fill in each column, starting each activity with the aim it serves.
Nonprofit program logic model
The W.K. Kellogg format most foundations and government funders ask for.
Starts from
The problem or need your program addresses
Columns
- Inputs
- Activities
- Outputs
- Short-term outcomes
- Mid-term outcomes
- Long-term outcomes
- Impact
GrantCopilot drafts the columns from your problem statement and linked proposal.
Logic model examples
An illustrative Theory of Change model for a broadening participation project. Read it left to right: every outcome follows from an output, and every output from an activity.
Goal
Increase the number of first-generation students who complete undergraduate research in environmental engineering.
- Inputs
- Faculty mentors
- Department and college support
- Two community college partners
- Prior pilot data
- Activities
- Recruit community college transfer students
- Develop a mentored summer research curriculum
- Train faculty and graduate student mentors
- Establish a peer cohort that meets monthly
- Outputs
Counted
- Students enrolled
- Mentors trained
Created
- Mentoring manual
- Research curriculum
- Short-term outcomes
- Participants gain research skills
- Participants report a stronger sense of belonging
- Mid-term outcomes
- Participants stay in engineering majors
- Partner colleges adopt the curriculum
- Long-term outcomes
- A more diverse environmental engineering workforce
- A reusable model for other institutions
An illustrative program logic model for an R25 research education program. Each activity names the Specific Aim it serves, so reviewers can trace every outcome back to an aim.
Program goal
Prepare undergraduates from underrepresented backgrounds for careers in cancer research.
- Specific Aims
- Aim 1: Give undergraduates hands-on cancer research experience
- Aim 2: Build the mentoring skills of faculty and postdocs
- Inputs
- Cancer center research labs
- Faculty and postdoc mentors
- Partner minority-serving institutions
- Activities
- Aim 1: Run a 10-week summer research program
- Aim 1: Hold weekly skills workshops
- Aim 2: Train mentors each spring
- Outputs
- Trainees enrolled each summer
- Workshops delivered
- Mentors trained
- Trainee posters presented
- Outcomes
Short-term
- Trainees gain lab and communication skills
Mid-term
- Trainees apply to graduate or medical school
Long-term
- Alumni enter biomedical research careers
- Impact
- A larger, more diverse cancer research workforce
An illustrative W.K. Kellogg style model for a youth literacy program. Outcomes move from knowledge, to behavior, to condition, the order most foundations expect.
Problem or need
Many third graders in our county read below grade level, and families have few free tutoring options.
- Inputs
- Program staff
- Trained volunteer tutors
- Library meeting space
- Donated books
- Activities
- Recruit and train volunteer tutors
- Run twice-weekly after-school tutoring
- Hold monthly family reading nights
- Outputs
- Students tutored
- Tutoring sessions held
- Books sent home
- Short-term outcomes
- Students gain reading skills and confidence
- Parents learn ways to read with their kids
- Mid-term outcomes
- Students read at home more often
- Families attend library events
- Long-term outcomes
- More students read at grade level by fourth grade
- Impact
- Higher literacy rates across the county
How the logic model builder works
Step 1
Pick your funder
Choose NSF, NIH, or nonprofit. Each one gets the columns that funder expects, so you are not reshaping a generic template.
Step 2
Start from your own words
Enter your starting point: a goal and research questions, a program goal and Specific Aims, or the problem your program addresses.
Step 3
Draft and edit
For NSF and nonprofit models, GrantCopilot drafts the columns from what you wrote. It does not invent partners, numbers, or institutions. Every entry is editable.
Step 4
Link it to your proposals
Link a logic model to one or more proposals. It exports as a table in their PDF and Word documents, and AI review checks your draft against it.
Learn the terms in the logic model glossary entry.
Frequently asked questions
Does NSF require a logic model?
The PAPPG does not require a logic model; individual solicitations decide. S-STEM, for example, requires a logic model or theory of change, ATE makes one optional, and EPSCoR RII-FEC accepts one. When it is included, it usually goes in the Project Description and counts toward your page limit, though S-STEM asks for it as a Supplementary Document. Read your solicitation, and ask your Program Officer if it is unclear.
Does NIH require a logic model?
Only when the funding announcement asks for one. The parent K12 announcement, for example, expects evaluation plans to include the program logic model, and some center and intervention research announcements ask for or strongly suggest one. The parent R25 announcements require an evaluation plan but do not mention a logic model, though one is a practical way to build that plan. Research grants such as R01 and R21 rarely need one unless the announcement says so. Check the announcement you are applying to.
Do nonprofit funders require a logic model?
Some foundations require one, and many more ask for its pieces: goals, outcomes, and an evaluation plan. Government funders vary: some require one and score it, and some publish their own logic model and ask you to show how your work fits it. The W.K. Kellogg format (inputs, activities, outputs, outcomes, and impact) is a common starting point. Even when it is optional, it gives the reader a one-page view of how your program works.
What is the difference between an output and an outcome?
Outputs are the direct products of your activities, such as the number of people served or a curriculum you wrote. Outcomes are the changes that follow, such as new skills, changed behavior, or improved conditions. Every outcome should follow from an output, and every output from an activity.
What is the difference between a logic model and a theory of change?
A theory of change explains why you expect your activities to produce change. A logic model lays out the chain from inputs through activities and outputs to outcomes, usually as a one-page visual. NSF's Theory of Change logic model combines the two: the chain plus the reasoning behind each link.
Will GrantCopilot make up details for my logic model?
No. A draft is built from your own text and the text of a linked proposal. Where your text says nothing about a column, entries are described by type, such as "Program staff" or "Faculty with expertise in the study area", and GrantCopilot does not name people, partners, or institutions or add numbers you did not give. You review and edit every entry before it goes into a proposal.
Which plans include logic models?
Every plan includes logic models, in the formats that match it. Planning and Professional include all three formats: NSF, NIH, and nonprofit. Nonprofit includes the nonprofit format, and Research includes the NSF and NIH formats. Plans start at $25/month. Every plan starts with a 7-day free trial.