Abstract
Within the literature on the evaluation of health (policy) interventions, complexity is a much-debated issue. In particular, many claim that so-called 'complex interventions' pose different challenges to evaluation studies than apparently 'simple interventions' do. Distinct ways of doing evaluation entail particular ontologies and epistemologies of complexity. They differ in terms of whether they define complexity as a quantitative trait of interventions, whether they see evaluation as part of or outside the intervention, and whether complexity can be regarded as an emergent property of the intervention and its evaluation. In practice, evaluators and commissioners of large health care improvement programmes rely on different, sometimes contradictory, repertoires about what it means to conduct a 'good' evaluation. This is an ongoing matter negotiated between and among commissioners, researchers, and-sometimes-programme managers. In particular, notions of evaluability, usefulness and distance/independence are problematised in different ways and with diverse consequences, which, in turn, produce other notions and layers of complexity such as temporal, institutional and affective complexities. When (social science) researchers claim that one method or another is better able to grasp complexity, they elide the issue that any methodological choice emphasises some complexities and lets others fade into the background. Analysing the practicalities and emotions involved in evaluation studies opens up the notion of complexity to analytical scrutiny, and suggests a basis for co-theorising between biomedical, public health and social scientists (including Science and Technology Studies scholars).