Logic Model

An articulated model of how a program or project is understood or intended to contribute to its specified outcomes and that focuses on intermediate outcomes rather than tightly specified processes is a logic model. Such models are usually shown diagrammatically but can be reported in narrative form. Although logic models have become increasingly popular in recent years, their use can be traced back at least to the 1960s, when Suchman suggested that evaluation might address the achievement of a “chain of objectives.”

Logic models can be developed prospectively for planned new programs or retrospectively for existing programs. Logic models can be used in various ways: (a) to guide an evaluation; (b) to provide staff and other stakeholders with a common, motivating vision; and (c) to report a performance story to funders and senior decision makers. Ideally, a performance story based on a logic model does not simply report evidence to support the causal chain but addresses the extent to which the outcomes can reasonably be attributed to the program and the influence of external factors.

Logic models can be drawn with the causal sequence going from left to right, from top to bottom, from bottom to top, or even from the outside into the center of a circle.

Where there are multiple strands of a program logic, it can be important to make it clear whether these are alternative causal paths or complementary paths. For example, in Figure 1, the program logic diagram for a home visiting service for parents of young children, the two strands are complementary—it is understood that both are needed for the program to succeed—which has implications for both practice and evaluation.

Nurses need to balance their advice on infant nutrition with a concern not to undermine parental confidence. Evaluation of such programs needs to focus on both these intermediate outcomes. Logic models that show such complementary strands can identify points at which programs need to achieve a reasonable balance between competing imperatives rather than optimizing one of them at the cost of the other.

It can also be important for practice and for evaluation to identify alternative causal paths—different ways in which participants might achieve the intended outcomes. These can be particularly important in making valid comparisons between the outcomes of participants in the program and of nonparticipants. It cannot be assumed that the program is the only way in which the intended outcomes can be achieved or that the program works in the same way for all participants. In some cases, it is possible to show the particular contexts in which particular causal paths will operate, taking a realist approach to thinking about programs and evaluations.

There is some debate as to whether a diagram of boxes labeled Inputs, Processes and activities, Outputs and short-term outcomes, and Outcomes and long-term outcomes (sometimes preceded by a fifth box, labeled Context) is sufficient to constitute a logic model (see Figure 2 for an example used by the United Way). [Page 233]

Figure 1Complementary Paths in a Logic Model

Figure 2Illustration of a Simple Logic Model

Reprinted from Measuring program outcomes: A practical approach (1996). Used by permission, United Way of America.

Figure 3Representing Other Factors in a Logic Model

From Halpern, G., “From hubris to reality: Evaluating innovative programs in public institutions,” in The Innovation Journal, 3(3), copyright © 1998. Reprinted with permission.

...