Table 1.

Philosophical assumptions underpinning approaches to research impact

Perspective	Positivist	Constructivist	Realist	Critical	Performative
Assumptions about what [research] knowledge is	Facts (especially statements on relationships between variables), independent of researchers and transferable to new contexts	Explanations/interpretations of a situation or phenomenon, considering the historical, cultural and social context	Studies of how people interpret external reality, producing statements on ‘what works for whom in what circumstances’	Studies that reveal society’s inherent conflicts and injustices and give people the tools to challenge their oppression	Knowledge is brought into being and enacted in practice by actor-networks of people and technologies
Assumed purpose of research	Predictive generalisations (‘laws’)	Meaning: perhaps in a single, unique case	Theoretical generalisation (what tends to work and why)	Learning, emancipation, challenge	To map the changing dynamics of actor-networks
Preferred research methods	Hypothesis-testing; experiments; modelling and measurement	Naturalistic inquiry (i.e. in real-world conditions)	Predominantly naturalistic, may combine quantitative and qualitative data	Participatory [action] research	Naturalistic, with a focus on change over time and network [in]stability
Assumed way to achieve quality in research	Hierarchy of preferred study designs; standardised instruments to help eliminate bias	Reflexive theorising; consideration of multiple interpretations; dialogue and debate	Abduction (what kind of reasoning by human actors could explain these findings in this context?)	Measures to address power imbalances (ethos of democracy, conflict management); research capacity building in community partner(s)	Richness of description; plausible account of the network and how it changes over time
Assumed relationship between science and values	Science is inherently value-neutral (though research can be used for benign or malevolent motives)	Science can never be value-neutral; the researcher’s perspective must be made explicit	Facts are interpreted and used by people who bring particular values and views	Science must be understood in terms of what gave rise to it and the interests it serves	Controversial; arguably, Actor-Network Theory is consistent with a value-laden view of science
Assumed mechanism through which impact is achieved	Direct (new knowledge will influence practice and policy if the principles and methods of implementation science are followed)	Mainly indirect (e.g. via interaction/enlightenment of policymakers and influencing the ‘mindlines’ of clinicians)	Interaction between reasoning (of policymakers, practitioners, etc.) and resources available for implementing findings	Development of critical consciousness; partnership-building; lobbying; advocacy	‘Translations’ (stable changes in the actor-network), achieved by actors who mobilise other actors into new configurations
Implications for the study of research impact	‘Logic models’ will track how research findings (transferable facts about what works) are disseminated, taken up and used for societal benefit	Outcomes of social interventions are unpredictable; impact studies should focus on ‘activities and interactions’ to build relations with policymakers	Impact studies should address variability in uptake and use of research by exploring context-mechanism-outcome-impact configurations	Impact has a political dimension; research may challenge the status quo; some stakeholders stand to lose power, whereas others may gain	For research to have impact, a re-alignment of actors (human/technological) is needed; focus on the changing ‘actor-scenario’ and how this gets stabilised in the network