A Practical Application of the Need to Fill the Gap Between Theories of Action and the Implementation and Evaluation of a New Program or System
If you’ve ever worked with someone from the Center, been in a Center staff meeting, or even had dinner with someone from the Center, you know that we refer to Theories of Action incessantly. It may sound wonky and weedy (and it is), but there’s a reason why we value it so much. That’s because a theory of action (TOA) can help us clarify what we truly believe should happen if a program or system is implemented.
Defining a Theory of Action to Help Guide Longer-Term Goals
We at the Center have presented a definition of a theory of action, particularly as it relates to accountability system design and implementation (Marion, Lyons, & D’Brot, 2016) that can be applied to the design of any new system:
A theory of action outlines the components of the system, while clearly specifying the connections among these components.
Most importantly, a theory of action must be based on a specific goal or outcome while specifying the hypothesized mechanisms or processes for bringing about that intended goal. In the case of system or program design, a theory of action should describe how the particular clear goals will be achieved as a result of the proposed system(s). This allows designers to determine the processes that must be in place to achieve their goals and what empirical evidence must be collected to support those proposed goals and expectations.
But a theory of action isn’t enough. In a previous CenterLine post, I wrote about the value of evaluation from the onset of a program. I raise the need for evaluation because system designers need a plan of attack to help evaluate whether their theory of action holds once the system is implemented. Program evaluation is a systematic method for collecting, analyzing, and using information to answer questions about projects, policies, and programs, particularly about their effectiveness and efficiency (Shackman, 2018). Program evaluation as a process can be used to determine how well we can address a few key questions:
- Can we generalize our findings?
- Are we meant to generalize our findings?
- What can we learn about our program efforts and the impact of the program?
In this post, I offer the argument that an additional layer is needed; that a theory of action is insufficiently specific to organize an evaluation plan.
Leveraging Logic Models and a Theory of Action to Establish Your Program Evaluation
A logic model is a systematic and visual way to present and share your understanding of the relationships among the resources you have with which to operate your program, the activities you plan, and the changes or results you hope to achieve (Kellogg Foundation, 2004). It is intended to concretely articulate the strategies and activities that are necessary to actually achieve a series of outcomes that should be related to a larger theory of action. Without a specified logic model, you’re often shooting in the dark regarding what program or project success actually looks like.
A theory of action, coupled with a series of logic models that address each component of the TOA, can work together to clarify efficient evaluation practices for a system, program, or project. Specifically, logic models facilitate the ability to identify measures of success or progress that are needed to best understand whether the components of a logic model are being implemented. This information can be tracked through inputs, outputs, and outcomes (i.e., short, mid-, and long-term outcomes). Thus, strong program or project design allows us to use logic models to flesh out the theory of action. Since it’s winter, I’d like to raise an example that is a little more reminiscent of warmer days.
How I Applied a Logic Model to a Theory of Action to Complete a Home Improvement Project
I built a paver patio in the summer of 2018. It was equal parts fun and work. I had a very simple theory of action about how to improve my quality of life:
Although relatively straightforward, that theory of action is really a series of hopes that require a very clear plan of attack in order to achieve them. That’s where my logic model comes in–and note that a logic model has to align to a specific goal or outcome:
*Note: Due to the simplicity of the tasks, the short-term outcomes are not typically reflective of the coherent link between the output and the long-term outcomes.
For the sake of brevity, I have combined steps two and three from my theory of action above and described them in the logic model below, under the goal of Setting up the Patio to Party. Typically, I would suggest developing a logic model for each of the components in a theory of action. The table below describes the steps to create an outdoor oasis and facilitate good times with friends and family.
Now, there are a whole host of other things that are necessary to make this theory of action a reality. However, if I clearly identify the goals associated with each component of my theory of action, I can establish a logic model based on that component’s goal and evaluate whether I’ve actually pulled this off.
Ultimately, program, project, or system designers have an overall set of goals that they would like to implement. By establishing a reasonable theory of action, coupled with a clear set of logic models for each component of a TOA, it becomes much easier to evaluate whether we are meeting both the smaller and larger goals of the project.
Using Logic Models to Set Forth Changes in Educational Assessment and Accountability
Consider the case of educational initiatives. Many are complex, multi-faceted, and take a number of years to take effect. Even when implemented using a coherent logic model, we still face evaluation challenges because we need to be specific in what we evaluate. We need to avoid replicating the scenario used in a famous Sidney Harris (2019) cartoon that demonstrates our overreliance on miracles in our systems planning (as discussed recently at the 2018 Reidy Interactive Lecture Series on validation):
Coupling a theory of action with a series of logic models to support program monitoring and evaluation is a powerful step toward documenting the success of a system and its program. Moreover, a logic model can help us determine where certain components fell short, what components hit the mark, and identify specific areas of improvement for our next round of implementation.
Just in case you were wondering, people love my new patio. And when I’m home, it has improved my quality of life. Now it’s time to work on that logic model for getting travel under control so I can enjoy it more often!