ODI's Research and Policy in Development, Context, Evidence and Links Framework is an analytical tool to understand how research-based and other forms of evidence contribute to policy and practice. It is aimed at researchers, policy makers and practitioners involved in policy-relevant research rather than academic research. It can also be used as a practical tool to determine what needs to be done to influence specific policies and practices.

We define both research and policy very broadly. By research we do not just mean classical scientific research. It include any systematic learning process - from theory building and data collection to evaluation action research. Similarly, policy is not just narrowly defined as a set of policy documents or legislation; it is about setting a deliberate course of action and then implementing it. It includes the setting of policy agendas, official policy documents, legislation, changes in patterns of government spending to implement policies, and the whole process of implementation. It is also about what happens on the ground: a policy is worth nothing unless it results in actual change. These are all relevant if we want to try to make policy more evidence-based and see the results of our research adopted in policy and practice.

There is a vast amount of existing theory on this subject - you can read all about many of them on our website. But most of it is from developed, OECD countries and there is very little systematic research on the interface of research and policy in developing countries. This is a serious problem given the massive diversity of cultural, economic, and political contexts in the developing world - particularly given the weaker research and policymaking capabilities and democratic deficits that characterize some developing countries. Furthermore, international actors have an exaggerated impact on research and policy processes in the South. This makes it difficult to draw valid generalizations and lessons from existing experience and theory.

Some of the theory though does seem particularly relevant for developing countries. Roe identifies the importance of policy narratives. Policy makers are strongly influenced by very simple stories such as the 'tragedy of the commons', or that planting trees water attracts rain and preserves ground water. Many of these simple stories are wrong, but they are nevertheless very attractive and powerful. Lipsky points out the importance of street-level bureaucrats. It is the people who implement policy who very often have the greatest impact on how that policy translates into practice. Without understanding the policy implementation process and the people involved in it, it is impossible to know how to influence it to promote better policies and practice.

Malcolm Gladwell's book 'The Tipping Point' describes how social epidemics spread. It is about the different types of people who are involved in the policy process: connectors, who know a lot of people; mavens, who hoover up and digest information; and salesmen who are very good at 'selling' ideas. He describes research into US news anchors in the run-up to the elections in the United States, which showed how very small differences in the way they behave on screen can give very strong messages to the public. He talks about how the context affects how people behave. In another experiment in the US, researchers sent student on errands all over the campus, and arranged for them to pass somebody in distress who clearly needed help and anaysed the factors which influenced whether the students stopped to help or not. The most important factor seemed to be whether the student was in a hurry or not. He describes how some ideas seem to be "sticky" - the factors that determine whether people remember specific bits of information. Gladwell describes how the conjunction of these factors create the "tipping points" when ideas suddenly spread and are adopted.

Policy-making used to be widely thought of as a linear and logical process, in which policy-makers identified a problem, commissioned research, took note of the results and made sensible policies which were then implemented. Clearly that is not the case. Policy-making is a dynamic, complex, chaotic process. Clay and Schaffer's book 'Room for Manoeuvre' in 1984 described "the whole life of policy is a chaos of purposes and accidents. It is not at all a matter of the rational implementation of decisions through selected strategies". That is increasingly recognised as a more realistic description of the policy process than the linear rational model - though the truth is probably somewhere in the middle. Furthermore, as Steve Omamo pointed out in a recent report on policy research on African agriculture: "Most policy research on African agriculture is irrelevant to agricultural and overall economic policy in Africa". It is not really surprising that the link between research and policy is tenuous and difficult to understand if policy processes are complex and chaotic and much research is not very policy relevant.

The aim of our framework is to simplify the complexity of how evidence contributes to the policy process so that policy makers and researchers can make decisions about how they do their work to maximise the chance that policies are evidence-based, and that research does have a positive impact on policy and practice. It is based on a thorough review of the literature and a wide range of case studies at international, regional and national level across the developing world. It identifies four broad groups of factors. We call the first external influences. These are the factors outside a particular country which affect policy makers and policy processes within the country. Even in big countries such as India, international economic, trade and even cultural issues matter a great deal. In smaller, heavily indebted countries, World Bank and Bilateral Donor policies and practices can be very influential. At national level the factors fall into three main areas. The political context includes the people, institutions and processes involved in policy making. The evidence arena is about the type and quality of research and how it is communicated. The third arena links is about the mechanisms affecting how evidence gets into the policy process or not.

The results of our research indicate that The Political Context area is probably the most important. This includes issues about governance including democracy, the openness of policy process and media and academic freedom. Policy making is a very political process. If there is a large degree of contestation about an issue it is very difficult to make progress. By contrast, if there is a large demand, particularly spurred by a crisis where policy-makers are seeking a solution, the chances are much greater that research will be used. The process of policy implementation is also a crucial area. The street-level-bureaucrats who actually implement policy can exert an enormous influence. Despite what we said earlier, policy processes are clearly not completely chaotic. There are budget and legislative cycles which may be quite transparent and if researchers understand and feed their results into these cycles they have a much greater chance of success. Crises provide another entry point. In these situations, where the decision is likely to be taken very quickly, researchers need to respond very quickly and they are not always able to do so.

In terms of the evidence, there are a set of issues which seem to come out most clearly and which make a big difference to whether research is taken up. If you can provide a solution to a problem and are able to put these on the table, you stand a greater chance of being able to influence policy. There are number of dimensions to this. The first is relevance: working on an issue which is topical and relevant to policy-makers makes it easier for them to engage with the research. The second is operational usefulness: this is not just about producing research which is topically relevant, but about providing research which suggests how that policy-maker may do something differently in his work. The third issue is credibility: this is about not only the content of the message or the approach of the research, but about who is saying it and their recognised expertise. Communication is crucial in both directions: in terms of researchers listening to policy-makers and in terms of engaging policy-makers in the research right from the beginning and of keeping them involved or in touch with that process.

The issue of links or mechanisms for bridging research and policy is perhaps the area about which we know least and it is a very complex area. We know that issues of trust and legitimacy, networks and working groups, are important. But there are further questions about what makes a successful working group and when it draws up research and helps to bridge the gap between research and policy. This is a challenge which we still face and which our ongoing work will provide further insights on.

Finally, there are some large external influences which shape and affect all of these issues about how research and evidence is used. One of the biggest influences or incentives at the moment is the 'War on Terror', which is shaping the way that policy is being made all around the world. These huge global agendas affect policy and how research and evidence is being used. In Europe the issue of EU accession is providing significant incentives for these countries to call in and engage with their researchers to inform policy for joining the EU. In Africa, there are Poverty Reduction Strategy Papers (PRSPs) which provide incentives for local policy-makers to pull in information from their researchers. What donors do also matters. DFID is developing a new research strategy which is shifting the emphasis away from basic research towards policy research and mechanisms for bringing research into practice. My final point about external influence is that donors can have a mixed role. If large organisations like the International Monetary Fund seek to identify and impose their findings of what the key issues are on a recipient country, there is likely to be a considerable backlash. On the other hand, donors can add legitimacy to the findings of local research which were previously considered problematic.

The framework we present here is a generic, perhaps ideal, model. In some situations there will not be much overlap between the different spheres, in other cases the overlap may vary considerably. In situations where there is little political will for change, the context and evidence spheres may not overlap at all and intermediaries are essential to bring the evidence to the notice of policy makers (the daisy-chain model). The ivory tower model describes situations where university-based academic research takes place in complete isolation from the real world outside. Furthermore the relative importance of each of the spheres may be different in different situations, and may change over time. The framework should perhaps be viewed as a trio of floating spheres of variable size and degree of overlap.

An interesting thing about the framework is how well it maps onto real-life activities. The political context sphere maps onto politics and policy making, evidence onto the processes of research, learning and thinking, and links onto networking, the media and advocacy. Even the overlapping areas map onto recognisable activities. The intersection of the political context and evidence represents the process of policy analysis - the study of how to implement and the likely impact of specific policies. The overlap between evidence and links is the process of academic discourse through publications and conferences, and the area between links and political context is the world of campaigning and lobbying. The area in the middle - the bulls-eye - where convincing evidence providing a practical solution to a current policy problem, that is supported by and brought to the attention of policymakers by actors in all three areas is where there is likely to be the most immediate link between evidence and policy.

So, if you are a researcher, policy maker or development practitioner with the desire to promote a particular policy you need to know about:

• the external environment which might influence how people think or behave: who are the key external actors? what is their agenda? And how do they influence the political context?
• the political context you are working in: is there political interest in change? is there room for manoeuvre? how do policy makers perceive the problem?
  
• the evidence you have, or could get: is there enough of it? is it convincing? is it relevant? is it practically useful? are the concepts familiar or new? does it need re-packaging?
  
• and the links that exist to bring the evidence to the attention of policy makers: who are the key organisations and individuals? are there existing networks to use? What's the best way to transfer the information: face-to-face or through the media or campaigns?

For researchers wishing to influence policy and practice, understanding the context, evidence and links is just the first part of the process. Our case studies also identify a number of practical things that researchers need to do to influence policy and practice, and how to do it. In the political context arena you need to get to know the policymakers, identify friends and foes, prepare for regular policy opportunities and look out for policy windows. One of the best ways is to work with them through commissions, and establish an approach that combines a strategic focus on current issues with the ability to respond rapidly to unexpected opportunities. Make sure your evidence is credible. This has much more to do with your long term reputation than the scientific credibility of an individual piece of research. Provide practical solutions to policy problems in familiar language and concepts. Action-research using pilot projects to generate legitimacy seems to be particularly powerful. Make the most of the existing links by getting to know the other actors, working through existing networks and building coalitions and partnerships. Identify the key individuals who can help. You need people who can network with others, mavens to absorb and process information, and good salesmen who can convince the sceptics. You may also need to use informal "shadow networks" as well as more formal channels.

Doing all of these things requires a wide range of skills. Researchers who want to be good policy entrepreneurs also need to be:

• Storytellers: Practitioners, bureaucrats and policy-makers often articulate and make sense of complex realities through simple stories. Though sometimes profoundly misleading there is no doubt that narratives are incredibly powerful.
  
• Networkers: Policy-making usually takes place within communities of people who know each other and interact. If you want to influence policymakers, you need to join their networks.
  
• Engineers: There is often a huge gap between what politicians and policy-makers say they are doing and what actually happens on the ground. Researchers need to work not just with the senior level policy-makers, but also with the 'street-level bureaucrats'.
  
• Fixers: Policy making is essentially a political process. Although you don't need to be a Rasputin or Machiavelli, successful policy entrepreneurs need to know how to operate in a political environment - when to make your pitch, to whom and how.

Try ODI's Policy Entrepreneur Questionnaire to find out whether you tend to favour, or to avoid any of these activities. If so, you may need to develop new skills in these areas, or work with others who have these skills.

Influencing policy change is an art as much as a science, but there are a wide range of well known and often straightforward tools that can provide powerful insights and help to maximize your chances of impact on policy. We've already seen how ODI's RAPID Framework can help you to understand the context you are working in and how you could use the Policy Entrepreneur Questionnaire to figure out what you are good at. Other useful tools to help to understand the policy context include Stakeholder Analysis, Forcefield Analysis, Writeshops, Policy Mapping and Political Context Mapping. This is vital in terms of developing an influence strategy. There is a wide set of research tools - from case studies to action research - that can help generate new or better evidence to support your case. The key communications questions are: Who do I want to convince? What do I want them to do? What will convince them? What relevant material do I have? A SWOT analysis can help to focus a communications strategy on the key messages and targets, and using the media can help you to reach a wide audience. Many tools have also been developed by organisations involved in lobbying, advocacy and campaigning for pro-poor change.

You can find out much more about these and other tools on the RAPID Programme website at www.odi.org.uk/rapid. The site also contains annotated bibliographies, details of RAPID research and meetings and downloadable versions of all of RAPID's publications. If you don't have good access to the internet all of the material on the website plus video clips and presentations from RAPID meetings is available on a CD-ROM.

If you would like further information, or to request a CD-ROM, please feel free to contact us by e-mail at: rapid@odi.org.uk; or by telephone at +44 (0)207 922 0300 or fax at: +44 (0)207 922 0399; or by writing to The RAPID Programme, Overseas Development Institute, 111 Westminster Bridge Road, London, SE1 7JD, UK.