This week on BetterEvaluation we're presenting Questions and Answers about logic models. A logic model represents a program theory - how an intervention (such as a program, project or policy) is understood to contribute to its impacts.
How do you draw logic models? Do you have suggestions or good examples to share? Or do you have more questions about drawing logic models?
Our focus this week honors the legacy of Carol Weiss, a pioneer in theories of change, who passed away last week at the age of 86. The Boston Globe obituary summarizes her many achievements. Way back in 1972, in her book, "Evaluation research: methods for assessing program effectiveness", Carol set out one of the first examples of using program theory to improve evaluation - unpacking the different ways that a teacher home visiting program might work.
Carol's work showed that logic models don't always have to be in the form input->process->outputs->outcomes->impact. Instead she showed how useful it can be to focus on teasing out the different possible causal paths between program activities and its outcomes.
What should a logic model look like?
There are two broad options. Many logic models are in the 'pipeline' form of input->process->outputs->outcomes->impact. Here's a logic model in this form for a project which provides a computer summer school to disadvantaged high school students.
The advantage of this style of model is its consistency. It's easy for people unfamiliar with the project to read the logic model. But its form puts all the activities at the front of the model. If the project includes other activities after the summer school, these can be difficult to represent,
The alternative style is an outcomes hierarchy where every box in the diagram is in the form of a result. This makes more explicit the theory that underpins the program. But these types of diagrams need to be carefully designed to make them easy to read, since they do not have a standard format.
Who should be involved in drawing a logic model?
Logic models benefit from bringing together different knowledge and perspectives, so it can be useful to involve people with direct experience of a project, content experts, and outsiders who can ask 'dumb' questions about hidden assumptions.
How can you draw a logic model?
Manual methods can be useful when you want to make it easy for many people to be involved in the process. Sticky notes can be easily moved around on a flip chart page or a wall. At the end of the meeting, it can be helpful to take a photo of the diagram. (Thanks to Charles Gasper for this great tip from the AEA365 blog.) A "magic wall", made by spraying a plastic shower curtain with an adhesive, provides a logic model which is more secure. (Thanks to Jess Dart for this suggestion).
To draw up a logic model in a final computer form, you can use generic software such as Word, Excel, Visio or PowerPoint. An advantage of using PowerPoint is that it is easy to present a logic model by gradually building it up.
Or you can use specialist software, which can also be used in real time with the group developing the logic model. Here are three we've seen:
DoView can produce logic models in the form a pipeline or an outcomes hierarchy, and it can produce 'nested' logic models with an overview version with additional pages that drill down to detail. (The outcomes hierarchy model above was drawn in DoView). To share a DoView model with someone who does not have the software, you can export it as a PDF file.
This has an easy interface which prompts you to complete the different components of the logic model and then produces the overall logic model. But it is not flexible, so it is only useful if the style suits your needs.
This also produces a logic model in a particular style. The software is accessed through a monthly or annual fee depending on the number of maps produced and users.
Weiss, C. H. (1972). Evaluation research: Methods for assessing program effectiveness. Prentice-Hall. Englewood Cliffs, NJ.