Theory-based evaluation acknowledges that, intentionally or not, all programs depend on the beliefs influential stakeholders have about the causes and consequences of effective social action. These beliefs are what we call theories, and they guide us when we design, implement, and evaluate programs.
Theories live (imperfectly) in our minds. When we want to clarify them for ourselves or communicate them to others, we represent them as some combination of words and pictures. A popular representation is the ubiquitous logic model, which typically takes the form of box-and-arrow diagrams or relational matrices.
The common wisdom is that developing a logic model helps program staff and evaluators develop a better understanding of a program, which in turn leads to more effective action.
Not to put too fine a point on it, this last statement is a representation of a theory of logic models. I represented the theory with words, which have their limits, yet another form of representation might reveal, hide, or distort different aspects of the theory. In this case, my theory is simple and my representation is simple, so you quickly get the gist of my meaning. Simplicity has its virtues.
It also has its perils. A chief criticism of logic models is that they fail to promote effective action because they are vastly too simple to represent the complexity inherent in a program, its participants, or its social value. This criticism has become more vigorous over time and deserves attention. In considering it, however, I find myself drawn to the other side of the argument, not because I am especially wedded to logic models, but rather to defend the virtues of simplicity.
Simplicity has its partisans. I am more of a fair-weather friend. Like Alfred North Whitehead I seek simplicity, but I distrust it. I make no appeals to an elegant universe or parsimonious rules that govern all human action. Occam’s razor is double edged, and I have felt its bite often enough to handle it with care. However, I continue to favor simplicity for a simple reason: it is often useful.
When a theory is simple, it is alive. It can be remembered and repeated, all the while arousing debate and discussion. It can spread quickly through word of mouth, outpacing more traditional modes of diffusion such as reports and planning documents. It can lodge itself in the thinking of those who take action, helping them consistently handle what is routine and successfully improvise when faced with the unexpected. With simplicity comes use. With use comes the promise of utility.
How can the benefits of simplicity be obtained and its perils averted? I don’t believe anyone has a complete answer – certainly I don’t – but I can offer some thoughts (Dare I say theories?) that I keep in mind when I work with stakeholders to develop program theories and craft representations of them.
- Theories are about thinking, not inking: A representation of a theory is less important than the process that gives rise to it. The reason we undertake the process is to enhance stakeholders’ shared understanding of a program theory. If logic modeling can achieve this, it is a worthy pursuit. If not, try something else. Regardless of the process we choose, the simpler the theory and its representation, the more quickly and completely a shared understanding will emerge.
- Dissent is welcome: When stakeholders share an understanding of a theory, they grasp its meaning and implications in very similar ways. Some stakeholders may not endorse every part of a theory. Others may not believe any of it. However, because they share an understanding of it, stakeholders can discuss what they consider to be the theory’s shortcomings in a comprehensible and productive manner. At some point stakeholders need to commit to one or more theories and representations that they will use in various ways, but this does not require lockstep agreement. Over time it is dissent that keeps a theory fresh. Simple theories and representations help foster debates that enrich a shared understanding rather than prolong the debate over what the theory is.
- Representations communicate imperfectly: A representation will always be more meaningful to those who created it, so as creators we can be fooled into believing that our diagrams, text, and matrices communicate our meaning more clearly than they actually do. We can only communicate about theories through representations, so our representations should be accessible to our intended audience. To ensure that this is the case, we should test what others understand our representations to mean and modify them accordingly. Simpler theories and models lead to less ambiguous representations.
- Match the process to the people: Rather than teaching stakeholders to use one “best” process of developing a theory and crafting a representation, modify one or more processes to fit the stakeholders. Start by learning about the people with whom you will be working. What are they like? What do they need? How can you help them get what they need? Be ready to think outside of the box (and the box and arrows). Remember, you are building understanding, not indoctrinating stakeholders. If you develop a theory in way that stakeholders find too complex, even if it derives from the humble logic model, they will get very little out of the process.
Simplicity has its limits. The conceptual challenge for evaluators is to establish those limits. The practical challenge is to help stakeholders take effective action. And when it comes to action, simpler is (almost) always better.