MIXED METHODS AND CROSS-CUTTING APPROACHES
22 Contribution analysis
Thomas Delahais
Abstract
Contribution Analysis is a theory-based evaluative approach particularly suited to the evaluation of complex interventions. It consists of progressively formulating “contribution claims” in a process involving policy stakeholders, and then testing these hypotheses systematically using a variety of methods (which may be qualitative or mixed).
Keywords: Mixed methods, complex interventions, contribution claims, abductive approach, context, causal pathways, causal packages, narrative approach
I. What does this approach consist of?
Contribution analysis is a so-called theory-based evaluation approach[1] (TBE): it is organised around a process of 1) developing a set of hypotheses about the effects of an intervention being evaluated (how these effects are achieved, in which cases, why…) – known as the ‘theory of change’; 2) testing these hypotheses through the collection and analysis of empirical information; and 3) updating the original theory by indicating which hypotheses are verified.
Like Realist Evaluation or Process Tracing, for example, Contribution Analysis is part of the new generation of TBEs that emerged at the start of the 2000s (sometimes referred to as theory-based impact evaluations – TBIE). It considers the interventions being evaluated as complex objects in complex environments. Central to Contribution Analysis is the postulate that interventions do not intrinsically ‘work’; their success or failure always depends on a diversity of drivers and contexts, which the evaluation needs to document. This is in contrast with Counterfactual approaches, for instance, which aim at identifying “what works” in isolation from their context. But what distinguishes Contribution Analysis from other approaches is that it also rejects the idea that the role of evaluation is to establish impact irrefutably: in a complex context, its aim is not to prove the effects of interventions, but to reduce uncertainty about their contribution to any changes that have occurred. It is in fact uncertainty that can be considered to be detrimental to decision-making and policymaking more generally.
I.I. Theory-building
The whole process of Contribution Analysis thus consists of gradually reducing uncertainty about the effects of the intervention being evaluated. As with all TBEs, the first phase, theory-building, consists of asking a question about the cause-and-effect relationships that are to be investigated and developing causal assumptions in response to this question. The latter usually concerns the contributions of the intervention to the desired changes. Let us imagine a governmental plan to prevent or deal with sexual violence in higher education institutions. The question asked might be: “How has the plan contributed to the effective reduction of sexual violence and better management of its consequences?”
The level of violence and the responses provided in institutions,