Inferring policy from experiment

Inferring Policy from Experiment was the title of a workshop held at the University of Kent on 15 May. There were three speakers:

Nancy Cartwright spoke about Two Approaches to EBHP:¬†Intervention-centring,¬†Context-centring. Cartwright distinguished between intervention-centred policy, which focuses on establishing a correlation between an intervention and the outcomes of interest and which forms the basis of the ‘what works‘ initiatives, and context-centred policy, which focusses on establishing the mechanism by which a policy works. Both have their advantages and disadvantages, but Cartwright viewed the context-centred approach as more promising overall. She also talked about the use of ‘markers’ and ‘cautions’ as evidence. See the Slides and hear the Audio.


Sarah Wieten asked, What are pragmatic trials good for? She explained the sense in which certain trials are ‘pragmatic’ (according to one view, there are ten dimensions with respect to which a trial can be pragmatic). Most importantly, perhaps, these trials have less strict eligibility criteria, involve less clinical expertise in their execution, and exert less control of compliance. Wieten also assessed the purported advantages of pragmatic trials. While these trials are supposed to achieve a better balance between internal and external validity, she argued that this is not the case. For example, sometimes one can depart from a pragmatic trial to increase internal validity at no cost to external validity. See the Slides and hear the Audio.


Mike Kelly talked about Inferring policy from evidence?  The case of non-communicable disease and health inequalities in the UK. In the UK, health is strongly correlated with social status, and inequalities have increased over the years, irrespective of government. Kelly bemoaned the fact that the key driver of government policy has been a focus on the individual: advising individuals as to how much to eat, drink, smoke etc. This shifts the responsibility for health onto the individual and away from the government. Kelly argued that we need a more collective, sociological approach to tackling non-communicable disease. Little progress will be made if governments ignore the evidence and expertise, and continue to focus on the individual. See the Slides and hear the Audio.


Many thanks to Veli-Pekka Parkkinen for doing a great job in organising this workshop on behalf of the Centre for Reasoning at Kent. Forthcoming Centre for Reasoning events are listed here.

Jon Williamson

Leave a Reply

Your email address will not be published. Required fields are marked *