It is particularly helpful in complex evaluation situations to examine causal mechanisms. At the core of working with causal mechanisms is the systematic analysis of basic interrelations between (elements of) an intervention and its expected outcomes. In formative evaluations, mechanisms can provide a better understanding of the way in which complex programmes are implemented. In summative evaluations, analysing causal mechanisms enables the evaluator to arrive at more credible statements on the causal links between an intervention and the observed outcomes.
However, the concept of causal mechanisms is applied differently across evaluation approaches. Mechanisms are described as causal processes in process tracing, contribution analyses and other theory-of-change approaches. Applying a broad analytical focus, they describe the entire causal process between intervention and outcomes. In the realist evaluation approach, by contrast, evaluators use what we call context-mechanism-outcome configurations. In this case, the intervention triggers or hinders the behaviour-changing mechanism. Looking for behavioural and psychological explanations, statistical procedures such as the meditation analysis examine “social mechanisms”, which they model and test as intervening variables. Moreover, causal mechanisms are also relevant in experimental designs, which make it possible to investigate the impacts of different elements of an intervention by varying the treatment arms.