National and international research institutes and NGOs have a growing interest in structured and more transparent methods of priority setting. In practice, they increasingly face similar problems in priority setting. Aside from selecting and applying appropriate methods, they have to ensure that various stakeholders are well represented. This is crucial for the results and implementation of identified priorities.

Until the 1980s, priority setting for agriculture research involved less transparent and structured procedures. Due to privatization policy in the 1990s the budget for public research has decreased significantly. Additionally, there has been growing pressure to show research results to justify expenditure. Since then, administrators have increasingly faced more openly expressed, sometimes conflicting demands of producers, agri-business, consumers, scientists, donors and politicians. Hence, there is a demand for new methods to assist in priority setting.
While there is growing attention for priority setting procedures, agricultural research programmes are considering the integration of biotechnology in their research programmes. Although there are a few examples in priority setting in biotechnology, its relevance needs to be assessed in the broader context of agriculture research. Therefore, this article looks at priority setting in general agricultural research.
Literally, prioritizing is choosing what has to be done first. Resource allocation is the usual over-riding economic objective of priority setting. The political rationale of structured and transparent procedure(s) includes:

• consensus building within and amongst different institutes;
• enhancing accountability amongst policy makers, researchers and end-users;
• motivating people's performance by relating their inputs to project results;
• helping  donors in project assessment.

Top-down versus bottom-up
Priority setting procedures can be broadly classified into top-down and bottom-up approaches. The top-down approach is dominated by officials and experts based on government goals and technical information provided at the programme levels (research leaders, scientists). National policies are also greatly influenced by priorities and support from International Agriculture Research Centres (IARCs) and by donors. Generally, at top government levels, policies are oriented at funding of agricultural versus non-agricultural research. At ministry levels, strategies are decided by comparing the various agricultural research programmes as they contribute to national goals. At the bottom-end of the hierarchy, at programme levels, resources are allocated based on technical considerations. Ideally, key actors from the ministry and top management of the research institute should participate at the top and bottom levels of the hierarchy, to facilitate communication and for a good match of policy and technical considerations.
The bottom-up approach essentially involves farmers, together with scientists, in the priority setting of problems and solutions. Farmers' needs, knowledge and priorities are solicited to formulate research agendas and identify research priorities. In other methods such as Participatory Technology Development (PTD), farmer participation goes beyond diagnosis and priority setting. PTD includes farmer-led experiments, development and evaluation. Although bottom-up approaches originate from NGOs, the methodologies are now increasingly adapted by some NARCs and IARCs. Two examples are briefly described.
Participatory Rural Appraisal (PRA) is a compilation of semi-structured activities carried out by interdisciplinary teams in partnership with communities and their local leaders. Although PRA deals with general community development, it could be specifically designed for agriculture research. Farmers can improve problem diagnosis and orient research to local issues and circumstances. PRA's visual approaches, and the cross-checking with interdisciplinary teams and community members can be an effective tool for implementing participatory research and development. However, the approach can suffer from imbalance in representation..
The Interactive Bottom-Up (IBU) approach's (see the box in the article by Commandeur) main strength is the involvement of different actors (scientists, farmers, governments, NGOs, donor) to a series of dialogues to assess problems and prioritize solutions for small-scale farmers regarding biotechnological innovations. Enhancing dialogue amongst different institutions is still relatively pioneering work. The IBU approach utilizes interdisciplinary perspectives to technology assessment and development, and assesses the comparative advantage of biotechnology over other existing technologies.
However, IBU has its weaknesses. Firstly, although the IBU approach proposes some sort of continuing involvement of farmers in technology generation, this remains vague; both in concept and in practice. The IBU approach still differentiates between farmers (as a source of information) and scientists (as developers of technology). In other words, farmers communicate their needs to scientists and the scientists, develop solutions for farmers. Hence, farmers' involvement is largely of the contractual and consultative types of participation. For the IBU approach, this type of dialogue and engagement is dependent on the good-will of the scientists, as is true for most participatory research methodologies.
Secondly, the IBU approach, as it has been adapted by the Special Programme on Biotechnology of the Netherlands' Directorate General for International Cooperation (DGIS), faces difficulties on how to respond to farmers' problems that cannot or should not be solved by biotechnology. For example, problems which are related to natural resource management or community development. In this sense, institutional relations, to link other development problems and activities effectively, need to be strengthened.

Methods of research priority setting
Once research problems are identified, there are several formally structured methods to assist in assessing and prioritizing options. These methods primarily deal with measurements, and do not substitute for the actors' experiences and contextual judgement. The discussions within these procedures are important in making decisions. Most of the methods described below are more commonly used in top-down approaches, but can also be adapted to bottom-up approaches. For instance, scoring and economic surplus methods are increasingly used by some NGOs. In addition, there is a growing number of government institutes and IARCs that use the PRA matrix ranking.

• Scoring. This is one of the two most commonly used quantitative methods in agricultural research priority setting. This involves ranking a set of research projects according to multiple criteria. Ranking results from the total scores given to the projects.

• Economic surplus. This is the other most commonly used method. It is used to calculate internal rates of returns and net present values of benefits generated from agricultural research projects. It can be used to assess the benefits of research to specific commodities or to a broader research programme. It can also be used to estimate the distribution of research benefits to consumers and producers. The calculation of rates of return to research can be compared to alternative public investments.

• Simulation models. In these models, mathematical relationships among variables are exposed to different market scenarios to asses the best outcome. These models can incorporate many factors that affect research priorities such as multiple goals and socio-economic constraints. The major disadvantage is that it requires a large investment in facilities. Additionally, data requirements are more extensive than for other economic models.
• Analytical Hierarchy Process (AHP). This involves comparative judgement of alternatives by identifying different levels (hierarchy) in decision making (see also the article by Ghislain, Nelson and Walker). Such decision levels include the overall goals, criteria, and alternative sets of research projects. Pair-wise ranking of each levels are made, after which alternatives are ranked.

Issues in priority setting
Given the diversity of needs and capacity of each institutions, there is no prescribed single-model to use in priority setting. Users should choose the methodology which suits them and their objectives best. Moreover, some of the methodologies are not mutually exclusive and could complement each other. At times, top-down and bottom-up approaches can use over-lapping methods. For instance, PRA matrix ranking and economic surplus methods are used in both approaches. However, due to their complexity and costs, methods such as simulation models tend to exclude the users in bottom-up approaches. The effectiveness of priority setting is enhanced by the appropriateness of method(s) and essentially by how the methods are used.

• Representation. As other Monitor articles in this issue illustrate, representation in the priority setting exercise greatly influences the outcomes. Not only among, but also within target groups there is differentiation in interest. For example, better-off farmers with access to irrigation and better soil quality would naturally have different criteria in crop and varietal selection than resource-poor farmers managing different farming systems. In PRA, the problems and solutions selected within a community, are greatly dependent on the representation of the community members. PRA tools that rank households according to socio-economic status, and tools to assess gender activities may indicate but cannot address power relations. Nevertheless, if not considered, outside interventions can reinforce power relations. In terms of addressing social differentiation, PRA is not a substitute for community empowerment activities. In this sense, scientists could actively seek cooperation from other local institutions such as NGOs and farmers' organizations.
• Time, resources and impact. Priority setting is not a stand-alone activity. Inevitably, research activity is to be judged by its results; not solely by how well priorities are articulated. Priority setting exercises involve investment costs in human resources and expertise, data gathering and analysis. As the Chilean case suggests (see page 12), the information needs have to be assessed in terms of their usefulness to the exercise. For example, the time involved in the collection of a large amount of information may erode the relevance of the data. Moreover, participatory procedures could be tedious. In terms of cost-effectiveness, cost of farmers' involvement needs to be compared with for example, cost of technology research and development and the corresponding gains or losses when farmers adapt or reject the technology accordingly.

Editor, Biotechnology and Development Monitor

Sources
J. Alston, G.W. Norton and P.G. Pardey (1995), Science Under Scarcity: Principles and practice for agricultural research evaluation and priority setting. Ithaca and London: Cornell University Press.

J. Bunders and J. Broerse (eds.) (1991), Appropriate Biotechnology in Small-scale Agriculture: How to reorient research and development. Oxon: CAB International.

C. Falconi (1997), Methods for Assisting Priority Settings in Agricultural Biotechnology Research. Draft Document, ISNAR, The Hague.

G. Manicad (1996), A Training Module: Participatory research. ISNAR: The Hague.