Logic Models


The first thing you notice about graphic representations of logic models is that they often appear anything but logical. Or better to say that they tend to be kind of vague (but that may well be accurate since the logic of many programs is vague).

But it does not need to be that way. YOU can be the one who dazzles the rest by producing logic models that really do reveal the logic of how you expect something to work (hint: that's what it needs to do!).

In the abstract, a logic model is any representation (words, pictures, whatever) of how a thing (e.g., a program or intervention) accomplishes whatever it is that it is supposed to accomplish. How is it that doing X will make Y happen?


Figure 1. A logic model explains how we expect our intervention to change the world.1

Everyone longs for a blue world and they are enticed by your offer, but how, they wonder, will you do it? What do they (you) already know?


Figure 2. Take the world as it is, add resources, people, and an intervention (still unspecified) and produce a new world.2

Let's rest our eyes and minds on this equation for a moment. How does change happen? Our equation suggests bringing to "the world as we know it" (the world on the left) RESOURCES (symbolized here by a stack of Benjamins) and PEOPLE (the "us" here) and a certain black box called an "intervention" yields a new world (on the right).

OK, resources and people we can understand. But our funders, the general public, and, in particular, the people we plan to intervene on, would like to know more about that black box.

Opening the Black Box of an Intervention

Any attempt to intervene in the affairs of a community or organization involves, first and foremost, doing something: actual activity.

Activities are done by someone. Activity produces some outputs. Outputs may or may not lead to actual changes in the community or organization (we call these outcomes).


Figure 3. People use resources to do activity that produces outputs. IMPORTANT: an OUTPUT (what an activity yields) is not the same as an OUTCOME.3

But outputs are not outcomes! The most frequently missing link in describing the logic of an intervention is how the organizational outputs effect/cause/produce the desired change in the world (i.e., the reason we are doing this intervention to start with).


Figure X.

Sometimes this happens by simply conflating outputs and outcomes. The youth diversion program funds will be used, for example, to run 100 after school sessions for kids in the neighborhood. But 1000 session-seats is not an outcome, it is an output. The funding was about changing the life chances of these kids. That's the outcome.

To connect outputs to outcomes we usually need something like a "theory of change" — how do these outputs effect change so as to produce the desired outcomes. This logic is often found under the heading "rationale" and it can be thought of as wrapping around the whole program.

Related to the rationale are the assumptions that underly the program. What facts about the world, laws, statistics, fixed resources, etc. are assumed and necessary for program success?


Figure X.

Let's go back to that original purpose. Recall our first picture which showed "the world now" and "the world after we are done with it." Let's parse this starting point. We usually begin with an observation about the world — sometimes called a problem statement. An explicit description of a condition in the world that our program is designed to change. Then we have a concrete goal — what observable change do we actually hope to accomplish?
We add these two elements as overarching starting points across the top of the picture:


Figure X.

And so let's stick the pieces back together:


Figure X.

And finally, we recognize that no program happens in a bubble. There is always an wider world, a political, cultural, and economic environment that constrains what can be done and can even being moving faster or in a bigger way than a program and hence threaten to dominate and wipe out any salutary effects it might have.


Figure X.

A Few Real World Examples

Let's look at a few examples and ask ourselves what the various graphic elements represent. The first thing we will notice is that not all published logic models have all the components listed above. One reason is that there are lot of substandard examples out there. Another is that there is no fixed and firm official definition of what a logic model has to contain. It is a tool. It needs to do something for you in a given context. The most general description of that task is to make as explicit as possible the relationship between what you need, what you will do, what you expect to accomplish and how in the context of a realistic assessment of environmental resources and constraints.


Figure X. Parent Training Program Logic Model

This parent training logic model is pretty typical. Look first at the inputs; notice some vagueness and that some "inputs" sound like activities. Notice too that the authors have opted to distinguish short and long term outcomes rather than outputs and outcomes. Next we have a "version 2.0" diagram from same program.


Figure X. "Family Involvement Project" Logic Model

Or how about this one:








What is a logic model? A conceptual explanation of how a program, intervention, initiative, etc. has an impact on the world. In general, logic models are represented visually and so the term usually refers to a particular type of diagram.

Although a main feature of many, if not most, of the logic models you can find out there is that they do not appear all that logical, there has evolved a sort of standard way of thinking about them. Being systematic in how we construct logic models can help us make more sensible ones and it helps others figure out what we are trying to show.

The standard list of components for a logic model is

  1. Inputs/Resources: what do we have to work with? what tools? what resources? what organizations? what people? what will we "use up"? what do we have to compete with other programs for?
  2. Activities: what will actors do?
  3. Outputs: Activities produce something. Services that are delivered. Posters are put up. These may or may not have desired effect on the world, but they are the things produced by the activities carried out.
  4. Outcomes: what changes result from the above in the entities we have targeted with our program/intervention?
  5. Impact: also called long term outcomes, how will the world be different after we have finished?

A starting point for building a logic model can be to think about three questions:

  1. What are we starting with? (assumptions, inputs)
  2. What do we plan to do? (inputs, activities)
  3. What do we hope to achieve? (Outputs, Outcomes, Impact)


Some sources suggest including problem statements, rationales/assumptions, and external/environmental conditions and relabel impact as "goals" and graphically suggest the iterative quality of things by making the diagram a loop rather than simple left to right sequence:

NOTES on Inputs

  1. Consider both positive and negative.
  2. People, organizations, resources
  3. Don't forget
    1. Equipment
    2. Space
    3. Expertise
    4. Time
    5. Networks, social capital
    6. Law, regulation
  4. Risks
  5. Barriers, obstacles, opposition

NOTES on Activities

  1. What is the money being spent on?
  2. What do the people you will hire actually do? (account for everyone?)
  3. Select a single "level" of detail and apply across the board.

NOTES on Outputs, Outcomes, Impacts

  1. Do not confuse outputs with outcomes. Outputs are just whatever you produce. It may be useless but it is still an output. A public awareness campaign, for example, might design and produce and even post beautiful fliers and posters, but if no one reads them or heeds them or becomes more aware because of them then they might be outputs without outcomes.
  2. Outcomes and Impacts are the things we often try to detect and measure when we do [WWW]program evaluation.
  3. Outcomes (and even more so impacts) are what constitute the benefits that justify the costs.

NOTES on Assumptions, Environment, etc.

One reason for doing a logic model is to get ourselves to think systematically about all the things that can affect whether or not we are successful. The question of whether or not to include something should be made with that in mind. In theory there is a "grand logic model" that really captures everything out there that we need to be aware of, but we probably never create that model (one might say it's the reality that we are trying to model): our models will always be simplifications.

NOTES Overall

There is no THE LOGIC MODEL. Different models make sense at different stages of a project. Different logic models make sense for different audiences. What to include or leave out should be dictated by the purpose at hand. Sometimes it's communication, sometimes it's exploration, sometimes it's mental book-keeping (keeping track of all the different things folks have reminded us about).

student-logic-model-01.jpg student-logic-model-02.jpg Template from LogicModelGuidepdf.jpg


Readings and References

  1. W K Kellogg Foundation. Logic Model Development Guide
  2. Innovation Network. Logic Model Workbook
  3. Some Word & Excel Templates

Other Resources

  1. [WWW]http://www.hfrp.org/publications-resources/browse-our-publications/learning-from-logic-models-an-example-of-a-family-school-partnership-program
  2. Wikipedia article on logic models
  3. CDC Evaluation Guide
  4. W.K. Kellogg Foundation Logic Model Development Guide
  5. U Idaho Extension
  6. Download a Worksheet from Shaping Outcomes