It is difficult to feed research-based evidence into policy and practice. This article discusses which capacities need to be strengthened to increase the impact of research on policy.

The world is changing rapidly in ways that often affect poor countries most. Economic, climate and population changes over the coming decades will have enormous implications for the challenge of reducing poverty by threatening access to food and water, worsening migration pressures and possibly increasing the chances of conflict. New research is essential for finding ways to prevent or mitigate the impact of these changes.

Donors are already spending over US $2 billion annually on development-related research. Yet there is widespread recognition that research alone is not enough. For research to have any impact, the results must inform and shape policies and programmes, and be adopted into practice.

Research donors increasingly acknowledge this. The UK Department for International Development (DFID), for example, will double spending on development research from US $200 to $400 million per year over the next five years, and will invest equally in generating new knowledge and working to ensure it is used in policy and practice.

The challenge of maximising the impact of research on policy and practice is not unique to multilateral and bilateral donors. Civil society organisations in developed and developing countries are not only engaged in practical programmes delivering services and strengthening systems to combat poverty directly, but are increasingly engaged in work to foster better development policies and programmes. Effective use of research-based knowledge is vital for both tasks.

This article outlines why it is so difficult to get research-based evidence into policy and practice. It provides some examples of what seems to work, describes a practical approach to developing effective strategies and identifies some of the capacity issues that need to be addressed.

Why is it so difficult?

Research results often need to be contested, debated and tested again before a consensus can be reached on recommendations for policy and practice. Even then many obstacles remain. Policy processes are very rarely linear and logical. Simply presenting research results to policymakers and expecting them to put the evidence into practice is very unlikely to work. Although most policy processes do involve a sequence of stages from agenda-setting through decision-making to implementation and evaluation, they rarely take place in an orderly fashion. Many agents are involved in affecting the process directly, and in trying to influence each other. While the whole process of policy has been described as ‘a chaos of purposes and accidents', I prefer to use the terms complex, multifactorial and non-linear.

Research-based evidence often plays a very minor role in policy processes. A recent ODI study of factors influencing chronic poverty in Uganda found that only 2 of 25 were researchable issues. In a talk on evidence-based policymaking at ODI in 2003, Vincent Cable, a senior member of the UK parliament, said that politicians are practically incapable of using research-based evidence because, among other things, few are scientists, and they don't understand the concept of testing a hypothesis. In another ODI meeting, Phil Davies, then deputy director of the governmental and social research unit in the UK Cabinet Office, described how policymakers tend to be more heavily influenced by their own values, experience, expertise and judgement, the influence of lobbyists and pressure groups and pragmatism based on the amount of resources they have available rather than by research-based evidence. In developing country contexts, national policy processes are often distorted by international factors. Donor policies, for example, can be hugely influential in highly indebted countries.

Researchers wishing to maximise the impact of their work have to attract the interest of policymakers and practitioners and then convince them that a new policy or different approach is valuable, and foster the behavioural changes necessary to put them into practice.

What seems to work

Research-based evidence can contribute to policies and practices that have a dramatic impact on peoples' lives. One example is the Tanzania Essential Health Interventions Project (TEHIP), in which the results of household disease surveys were used to inform the development of health services focusing on the most common conditions, especially those affecting mothers and young children. This contributed to a 43 and 46% reduction in infant mortality in two districts of rural Tanzania between 2000 and 2003. Another example is the Decentralised Livestock Services in the Eastern Regions of Indonesia Project, in which a careful combination of pilot field-level projects, institutional research and proactive communication contributed to a 250% increase in farmer satisfaction with livestock services. Success stories quoted in DFID's new research strategy include a 22% reduction in neonatal mortality in Ghana by having women begin breastfeeding within the first hour after birth and a 43% reduction in deaths among HIV positive children using a commonly available antibiotic.

These and other case studies from around the world illustrate the complexity of engaging with policy processes. There is no simple blueprint for what will work. What works in one context may not work in another. But it does appear that research projects and programmes are more likely to be successful when they:

• focus on current policy problems and have clear objectives;
• engage closely with policymakers throughout the process, from identifying the problem, undertaking the research itself and drawing out recommendations for policy and practice from the results;
• understand the political factors which may enhance or impede uptake and develop appropriate strategies to address them;
• invest heavily in communication and engagement activities as well as the research itself and build strong relationships with key stakeholders.

Individual champions and opponents frequently play a major role, as does serendipity - or chance.

The implications of this are that engaging with policy requires more than just research skills. According to Simon Maxwell, director of ODI, if researchers want to be good policy entrepreneurs, they also need to synthesise simple, compelling stories from the results of the research. They need to be good networkers to work effectively with all the other stakeholders involved in the process, good engineers to build programmes that can generate convincing evidence at the right time and political ‘fixers' who know who is making the decision and how to get to them. Or they need to work in multidisciplinary teams with others who have these skills.