aes17

Andrew Hawkins (ARTD Consultants)

Program logic is often conceived of as a model of a program's theory of change. This presentation argues that while logic and theory are very useful in evaluation, conflating the two leads to under-developed theory and over-extended logic. 

The presentation will start with a brief discussion of program theory, theories of change, and causal mechanisms. A theory may be as simple as a few sentences describing how and why a program is intended to work, or as in-depth as a collection of realist context-mechanism-outcome configurations. They differ in their level granularity. They share the common function of explaining why we expect change to occur. 

We will then visit the philosophy of causality to argue that program logic comprised of a series of boxes and arrows falsely implies that A leads to, or generates B (e.g. awareness of the program does not lead to participation in the program). Instead we will propose that a configurationist approach to causality with necessary and sufficient conditions provides a more useful underpinning to program logic than one implying a sequence of cause and effects.

The first step in intervention logic is diagramming the conditions considered necessary for achieving program outputs, and the assumptions we are relying on for this to hold. The second is diagramming the conditions that these outputs and assumptions must be sufficient for generating. The third is diagramming the conditions or outcomes these outputs should contribute to, along with the effects of external factors. By giving explicit attention to the interplay between aspects the program has relative control over (inputs and activities) and those that it does not (assumptions, external factors) this approach provides a holistic model of how an intervention is intended to be effective. Please bring a program logic you have developed to the presentation.