Our Programmes



Sign up to our newsletter.

Follow ODI

The role of research

Time (GMT +01) 11:00 12:30


Paul Spray – Director of Research, Policy Division, DFID
John Young – Research Fellow, ODI

Diane Stone - CSGR, University of Warwick

  1. Diane Stone introduced the third meeting in the series 'Does Evidence Matter?'. This meeting focused on the role of research, raising questions such as: What does DFID want from research? Does it get it? How could research have more impact on policy? How can you measure research impact? What is its relative importance in recent development policy shifts?
  2. Paul Spray spoke about the role and use of research in DFID. He started by pointing out that DFID distinguishes between 'research' and 'policy analysis'. While policy analysis is short term and always geared towards issues that DFID wants to know about quickly, research can be longer term and does not have to relate directly to DFID's work.
  3. However, all research that is carried out and funded by DFID has to contribute to the larger objective of poverty reduction. When choosing which research to fund, that the key criterion is how l plausibly the research will promote poverty reduction; that inevitably puts us at the applied end of research Research often shouldhave an element of challenge in it, perhaps challenging shorter term policy analysis.
  4. We don't know overall what the impact of research on poverty reduction is. However, we know that the link between research and poverty reduction does not necessarily have to go through government policy. We also know that policy decisions are not necessarily informed by research - even though they may give the impression of being based on rigorous facts.
  5. When are researchers influential? Researchers can seek to influence through snuggling up to policy makers. But they can also have an impact through being confrontational and contributing to conflict in a specific field - which forces policy makers to reflect on what is going on and to respond.
  6. It is important for researchers to catch the right moment in a policy process - e.g. when new ministers are appointed…At the same time, researchers can also be influential when they tackle emerging issues: governments need to know if there are big problems or opportunities ahed e.g. DEFRA funds 'horizon scanning' research.
  7. Research is widely popular with government when it produces a quick fix. Not all research should aim to do this. But it is worth noting that research is usually more effective when it is problem focused and when it is easy to measure its effects.
  8. There are a number of key issues that DFID is interested in at the moment. DFID needs to find out what its own niche is, both in international development research and in national research (e.g. in relation to ESRC). DFID London also needs to find out how central programmes can best relate to DFID country offices, especially in capacity building of southern research. There are issues to be worked out regarding how to leverage private sector research, how to disseminate research findings effectively and manage the relationship between researchers and users. In this respect, DFID may perhaps draw more on its convening power and stage dissemination events and workshops.
  9. DFID Policy Division has now moved to a more think tank type structure, with shorter term teams grouped around topical issues. Hopefully the team structure will enable DFID to draw in outsiders to work for shorter periods on different teams: this is an important opportunity which academics should seize
  10. John Young presented one aspect of the RAPID programme at ODI - how can researchers achieve greater policy influence. Clearly evidence can matter, but there is no systematic understanding of how and when it does. ODI has been examining this issue in a systematic manner since 1999, when Rebecca Sutton published her paper on the policy process.
  11. Traditionally the link between research and policy has been viewed as a linear process. This is clearly not the reality. Opinion is now shifting towards a more dynamic and complex model of research-policy linkages. To illustrate this, John presented two quotes on policy and research respectively: Clay & Schaffer (1984) say about policy that 'The whole life of policy is a chaos of purposes and accidents', and Omamo (2003) says about research that 'Most policy research on African agriculture is irrelevant to agricultural and overall economic policy'.
  12. RAPID uses a three-dimensional framework to understand research-policy linkages: political context, evidence, and links. The framework has been applied to four examples of policy change in four in-depth case studies: the adoption of PRSPs, the launch of SPHERE in the humanitarian sector, the (non)-evolution of animal health policies in Kenya, and the incorporation of Sustainable Livelihoods principles in the DFID 1997 White Paper. In addition, RAPID draws on evidence from the fifty summary case studies collected by the Global Development Network (GDN).
  13. What have we learnt from this material about what researchers need to do? In relation to the political context, researchers must get to know the policy makers, identify potential supporters and opponents, prepare for regular policy opportunities, and react to unexpected policy windows. Examples from the study on animal health policy in Kenya and the study on the Sustainable Livelihoods approach illustrate this.
  14. In relation to the 'evidence' dimension of the framework, researchers need to establish credibility and legitimacy, provide practical solutions to problems, and communicate effectively. For example, the Rwanda evaluation that led to SPHERE was influential largely because it was regarded as rigorous and credible. Also, action research and pilot projects seem to be effective means of convincing policy makers. One example of this is the influence of the PEAP programme in Uganda prior to the full adoption of PRSPs.
  15. In relation to 'links', researchers must get to know the other stakeholders, establish a presence in existing networks, build coalitions, identify key networkers and salesmen, and use informal contacts. Again these points can be illustrated with examples from the RAPID case studies.
  16. In conclusion, think-tank/do-tank/operational organisations appear to have more immediate policy impact than academic research institutes. However, academic research and 'free thought' contributes to the discourse in which policy is made.
  17. Comments from the audience focused on a range of issues:
  • How do we build constituencies at the national level to link research to the democracy/governance agendas?
  • The National Systems of Innovation (NSI) literature provides good insight into why some research is taken up and other research is ignored.
  • The quote from Clay & Schaffer ('policy is a chaos of purposes and accidents') would seem to suggest that research-policy linkages are far less rational than implies in the diagrams used by John in his presentation (on what researchers should do and how they should do it).
  • Researchers still work under conservative career conditions - e.g. they need to publish articles in peer reviewed journals. This may hinder inter-disciplinary work.
  • Will DFID collaborate with civil society organisations on research issues?
  • The sole objective of poverty reduction may limit the nature of development research.
  • The meeting has presented material that challenges researchers to behave differently. IT also challenges research funders to behave differently, e.g. by providing more funding for networks, dissemination and impact work. This is particularly important in relation to building research capacity in developing countries.
  • The sole aim of poverty reduction may exclude research within the humanitarian field.
  • Is all research really pro bono and therapeutic? There seems to be an underlying assumption in the meeting that all researchers are virtuous.


The third meeting in the series 'Does Evidence Matter?' focused on the role of research in development policy.

It raised key questions for discussion such as:

  • What does DFID want from research?
  • Does it get it?
  • How could research have more impact on policy?
  • How can you measure research impact?
  • What is its relative importance in recent development policy shifts?