5. Co-designing the data collection and sharing system

5.1 Meeting Information Needs

Co-managers will have overlapping data needs to support their roles (Figure 12). For example, many of the data types required for policy and development planning and evaluation may be similar to those required by local managers to evaluate the performance of their local management plans. The greater the overlap the more opportunities will exist to share data and information and the responsibility for collecting it.

Key stages in designing an effective and efficient data collection and sharing system are therefore identifying and maximizing this overlap, and reaching agreement on who should collect and share data to generate this information based upon their capacity and motivation. Responsibility for collecting the remaining data will also have to be reached. In some cases, the Government and LMI may be happy to collect these data independently of one another (often informally in the case of the LMI) but then later share them with one another. In other cases they may agree to collect data on behalf of one another provided they are sufficiently motivated to do so.

This chapter describes a participatory process for undertaking these key and other important stages when designing a data collection and sharing system for co-managed fisheries. The key elements of this chapter also form the basis of the accompanying Part1 of this paper: Practical guide.

FIGURE 12
Common data needs of co-managers

This design process should be integral to the formulation or revision of the management plan. It is at this stage when the overlapping data needs of important stakeholders' will become evident having established or clarified their roles and responsibilities, objectives, planned activities or strategies, rules and regulations, and performance evaluation criteria.

Participatory (and supporting conventional) monitoring programmes can subsequently be designed to meet these needs following an examination of the incentives and capacity of each stakeholder to collect and share the data, alongside information relating to the operational characteristics of the fishery. Integrating the design of the system with the management plan formulation, formalisation or revision process also provides opportunities to revise or re-negotiated iteratively elements of the management plan and/or the data collection and sharing strategy as the potential scope of any monitoring and evaluation activities becomes evident. Relevant activities required for formulating or reviewing management plans are therefore included in the eight-stage design process we propose in below.

5.2 AN EIGHT-STAGE PARTICIPATORY DESIGN PROCESS

The eight-stage process we propose below is based around a general strategy of identifying stakeholders and their information requirements, identifying overlap in these requirements, and then designing a strategy for data collection and sharing based upon stakeholder incentives and capacity (Figure 13).

The first step is to identify the main stakeholders involved in management of the resource, and their responsibilities and capacities (Stage 1) which will help define their potential roles in the system. Management planning (Stage 2) is key to designing successful data collection and sharing systems because the management objectives and strategies will be defined in the plan. Based on the objectives in the plan, the data that will need to be collected can be identified (Stage 3). Existing data that are already collected by different institutions are then reviewed (Stage 4) and gaps are identified, so that for the remaining data that are required by the stakeholders, a strategy can be identified to collect those data (Stage 5). Pathways and methods to share those data between stakeholders are agreed in Stage 6, and ways of recording, storing and managing the data are identified in Stage 7. Finally in Stage 8, the system is implemented, evaluated and refined. A scaled-down pilot system could be implemented at first, involving a reduced number of data variables and stakeholders, so that all involved can get a feel for the system and which ideas will work well or not in practice.

The process is likely to require the participation of the main stakeholders in a series of focus group discussions and planning workshops, possibly at different spatial or administrative levels, for example, district, regional, national, depending upon the administrative structures of the co-management institutions. The location and timing of meetings, the facilities where the meetings are held, and facilitation should be arranged to make sure that all stakeholders are able to participate in the process in ways that are important to them. Berkes et al. (2001) provide useful guidance in respect to planning field working activities of this nature.

Facilitators can help facilitate the design process. They can help organize meetings or workshops, bring stakeholders together and help lead the discussions and drive the process forward. Facilitators may be external - either independent facilitators or from an outside organization not otherwise involved in the data collection and sharing system, or they may be from an organization involved in the process, for example, from the lead organization, such as the Fisheries Department, local government office or an NGO or project.

Each of stage in the process is described below, with tools and ideas for how to facilitate each part.

FIGURE 13
A proposed eight-stage process for designing data collection and sharing systems for co-managed fisheries. The light green ovals in the diagram represent possible tools that can help carry out the stages.

5.2.1 Stage 1 - Identify the main stakeholders, their interest and capacity

The main purposes of this exercise are to:

• Ensure the participation of key stakeholders in the management planning process and/or design of the data collection and sharing systems. This has several benefits. Participating stakeholders will understand why information is being collected promoting effective participation. Participation in the design process will also ensure data collection systems are both practical and understandable and will increase a sense of ownership in the data generated and the management process. Both of these aspects will improve the quality of the data collected and the interest in it (Garaway and Arthur, 2002).
• Determine the nature of their interest in the resource, and capacity to monitor, evaluate and manage in preparation to formulating or formalizing the management plan and co-designing the data collection and sharing strategy (see below). Here stakeholder capacity includes resources (people and money), knowledge and skills, legal rights and motivation.

Stakeholder analysis provides a systematic means of identifying these key stakeholders and is the starting point of most participatory work and/or social assessments. As well as providing an understanding of power relationships, influence and interests of stakeholders, it can provide essential information about who will be affected (positively and negatively) by selected management strategies or interventions (see below), which individuals, groups, or agencies need to be involved in the management, and how; and whose capacity needs to be built to enable them to participate. Stakeholder analysis might also be used to identify current systems of information flow, where constraints lie and where new linkages can be made (see later).

Even if the co-management arrangements are already well established and management plans already formulated, this exercise may still be useful to clarify these issues or update knowledge and understanding in preparation for designing the data collection and sharing system.

Practical guidance for conducting a stakeholder analysis is provided in Part I of these Guidelines. Further guidance can be found at http://www.iied.org/forestry/tools/four.html or Annex D of IFAD (2002) http://www.ifad.org/evaluation/guide/index.htm. Using these techniques it should be possible to identify the main stakeholders belonging to the three main categories described in 2.3.1 as well as specific information on these groups, their interests and relationships. Depending upon the local context, stakeholders belonging to other categories or administrative levels of government may also be identified and described.

5.2.2 Stage 2 - Formulate, formalize or review local management plans

Properly formulated and clearly recorded management plans will greatly aid the identification appropriate indicators and data variables for inclusion in monitoring programmes as well as appropriate sources and methods.

The information contained within the plan will also be required to help coordinate management plans and activities and thereby minimize conflict among local managers (Section 3.3.10) and may also provide a rich source of variables for explaining the performance of both local management activities (Section 3.5.2) and co-management policy (Section 3.5.8).

In spite of the importance of the management plan not only in terms of designing monitoring programmes but also for coordinating and evaluating management activities, few examples of local management plans were identified during the preparation of this document. It is therefore highly likely that it will be necessary to formulate or at least formalize the management plan.

Broadly speaking, the formulation of the plan (excluding the design and development of monitoring and data sharing activities or programmes) is likely to involve the following steps:

• Describe the resource, environment, fishery, fishers and other stakeholders (see Section 3.3) drawing upon the results of stakeholder analysis described above and any baseline data collection activities such as Sustainable Livelihoods Analysis (SLA), Participatory Rural Appraisal (PRA) and frame surveys described below.
• Select or identify local management objectives consistent with national policy. If these are not explicitly stated then it will be impossible to identify appropriate indicators to monitor management performance in relation to them. Co-managers should be clear what they mean by objectives. For example, banning destructive fishing practices might be seen by some as an objective, but actually this is a management measure often selected to pursue objectives such as improving yield or equity. Effectively establishing and enforcing such management measures could however be regarded as an intermediate objective (see Berkes et al., 2001).
• Select management strategies, rules and interventions to achieve the objectives that comply with national legislation. The results of the stakeholder analysis (see above) can also be used to predict the effects of management strategies on key stakeholders.
• Collectively agree the roles and responsibilities of each stakeholder to (help) implement and evaluate the management plan. This should take account of stakeholder capacity identified during the stakeholder analysis described above. Table 1 may provide a useful checklist for this purpose. When identifying roles it is important to precisely establish the area of competence, geographical area, fish resources and fisheries for which a given stakeholder is responsible (Cochrane, 2002). Remember, if policy and legislation permits or are still evolving, or if changes in stakeholder capacity have occurred, it may be possible to revise these respective roles, or share responsibility for them, in order to create or strengthen incentives that encourage participation in monitoring programmes and data sharing activities by important stakeholders (Section 5.2.5.5). Note that sufficient time and resources should be devoted to this exercise since, according to the results of research activities described in Preparation of this document in Part I, identifying stakeholders and clarifying their roles and responsibilities is a critical but often challenging activity (see Box 14).
• Collectively agree on surveillance and enforcement activities including who is responsible for doing what.
• Develop legal and policy framework for management if not already in place.

Guidance on how to undertake each step is given by Hoggarth et al. (1999; in press). Intermediaries may be able to help facilitate the formulation of the plan between the relevant administrative levels of government and the LMI. Visualization techniques including “cognitive mapping” may facilitate these processes (Berkes et al., 2001; Guijt et al., 1998). In Cambodia, the MRRF project uses SWOT (Strength, Weaknesses, Opportunities and Threats) analysis to review and adapt the management plan for the following year (MRC, 2004).

Much of the baseline data and information can be readily compiled using Rapid Rural Appraisal (RRA) (Pido et al. 1996) and PRA approaches (see http://www.iied.org/sarl/pla_notes/ or Annex D of IFAD (2002), http://www.ifad.org/evaluation/guide/index.htm and Maine et al. (1996)) involving the LMI and resource users, as they will often know what actions or management control measures would sustain local catches (e.g. establishing fish sanctuaries in particularly locations). Intermediaries such as NGOs, research institutions or donor projects may be an important source for the socio-economic and institutional baseline data and information. In addition, formal stock assessments or the application of analytical frameworks (Section 3.5) may be employed where feasible and appropriate to investigate the biological, social and economic implications of different harvesting strategies and control measures (and their combinations) designed to control fishing mortality, and interventions such as stocking or habitat rehabilitation programmes.

5.2.2.1 A sustainable livelihoods approach to management planning

Poverty alleviation or reduction is often the central theme of national policy and development plans, as well as local management plans. The Sustainable Livelihood Approach (SLA) offers an effective means of identifying interventions or strategies in support of these policies and plans. The Department for International Development (DFID) formally adopted the approach in 1997 and defines the concept of sustainable livelihoods as:

“…the capabilities, assets (including both material and social resources), and activities required for a means of living. A livelihood is sustainable when it can cope and recover from stresses and shocks and maintain or enhance its capabilities and assets both now and in the future, while not undermining the natural resource base”.

Several international development agencies are now applying a “livelihoods approach”, each applying a somewhat different version and for a range of different purposes (see Krantz [2001] for further details). Most approaches reject the usual sectoral entry point (fisheries, agriculture, water, etc.) and instead begin with an analysis of people's current livelihood systems to identify appropriate interventions. They also place their emphasis on involving people in the identification and implementation of activities (ibid).

The micro or meso scale of most co-management initiatives lends itself well to the application of the SLA. Not only can more effective management plans be formulated using this approach taking account of the importance and impact of different sectors, but such initiatives also provide ‘entry points' for other, (inter-) sectoral management and development interventions including poverty reduction strategies (PRS) (See Section 3.2.3.2).

As well as providing a means of identifying context appropriate strategies and interventions including both technical and institutional type, the application of the SLA may also help facilitate the design the data collection strategy by:

• Identifying important stakeholders.
• Identifying issues, constraints and opportunities.
• Identifying alternative livelihood opportunities.
• Helping appraise the fishery to determine appropriate data sources and collection methods.
• Generating information that should be included in the local management plan.

A central element of DFID's SL approach is the SL framework (Figure 14). The framework is not an exact model of reality, but an analytical structure or tool to improve understanding of livelihoods. It aims to provide a broad and systematic understanding of the various factors that constrain or enhance livelihood opportunities. It can be used in both planning new development or management activities and interventions, as well as assessing the contribution to livelihood sustainability made by existing activities. It is built around five principal categories of livelihood assets depicted as a pentagon. An important stage of the analysis is to identify people's access to different types of assets (physical, human, financial, natural and social) and their ability to put these to productive use. It also facilitates an understanding of how organizations, policies, institutions and cultural norms shape livelihoods by determining who gains access to different assets and defining the range of livelihood strategies available and adopted (ibid). It does not take a sectoral view but rather seeks to understand the contribution made by all sectors to the assets upon which people can draw from. In essence, it aims to encourage users to take a broad and systematic view of the factors that cause poverty, whether these be shocks or adverse trends, poorly functioning institutions and policies, or a basic lack of assets, and to investigate the relations between them.

FIGURE 14
The sustainable livelihoods framework

Source: http://www.livelihoods.org/info/info_guidancesheets.htm

In doing so, appropriate interventions can be identified on the basis of threats or constraints to livelihoods such as providing people with better access to assets or supporting more effective functioning of structures and processes (e.g. management institutions, policies etc) that influence access to assets and available livelihood strategies. A full description of the SL concept, framework and application is available at http://www.livelihoods.org/info/info_guidancesheets.html. Hoon et al. (1997) examines appropriate indicators for SL analyses.

Poverty and Development Planning Applications

Norton and Foster (2001) conclude that the asset/vulnerability component of the framework has considerable potential to improve the PRSP, particularly in terms of the diagnosis, the design of the strategy, and the monitoring framework (see Box 17). DFID (2001) advocates that these benefits also extend to other country level development strategies including CDF and NSSD (See Section 3.2.3). Section 6 of DFID's Guidance Sheets available at http://www.livelihoods.org/info/info_guidancesheets.html describes a host of other opportunities where the application of SL approach, particularly its core principles can complement and support country-level policy and development planning.

5.2.2.2 Participatory Action Plan Development (PAPD) approach

Participatory Action Plan Development (PAPD) is an approach to management planning that was developed by the Center for Natural Resource Studies (CNRS), and a team from Newcastle and Durham Universities (Barr and Dixon, 2001). The method has been developed in Bangladesh but has been adapted and tested in Viet Nam (see Sultana and Thompson, 2003; 2004). The planning approach involves holding a series of linked local workshops where different stakeholders and users of a floodplain participate separately and in plenary to develop a management plan for the common resources they use.

The key features of PAPD

Each category of stakeholders works separately to identify and rank their problems regarding natural resource management (NRM). Later, all stakeholder groups come together to jointly agree on the priority problems. The stakeholder groups separately analyze possible solutions and their impacts, before meeting in plenary to share their analysis and form a consensus on win-win solutions and actions. The participants prepare in more detail an action plan for natural resource management. PAPD is designed to encourage participants to express their views, while avoiding a process that is dominated by locally-powerful and vocal people, and to develop a shared framework of understanding about resource management. It is based on certain principles such as the desirability of consensus, the need for all stakeholders to be involved in the process, neutrality and the sharing of information. However, it does not focus on negotiation or resolving existing direct conflicts between two parties over resources. PAPD focuses on problems, needs and potential solutions that are shared, and the differences and similarities in views of stakeholder groups over them. Many methods like Participatory Rural Appraisal (PRA) aim to raise individual awareness of resource management problems; PAPD as a process raises collective awareness of the problems and leads towards collective action that can tackle them effectively. PAPD is designed to enable the voices of the disadvantaged and less powerful to also be heard.

Methodology for PAPD

As originally conceived, PAPD was seen as a two-stage process comprising a problem census (listing of problems and ranking of their importance by different stakeholder groups) followed by stakeholder and plenary planning workshops. However, through the process of applying and testing the PAPD process, it has evolved into three phases that lead into continual or long-term participatory resource management (Sultana and Thompson, 2004). Each phase has a number of different stages and activities (Figure15).

- Scoping phase

1. Situational analysis (through summarizing local knowledge).
2. Stakeholder analysis (with help of key informants).
3. Household census and invitations to a random sample of households to PAPD (stratified by stakeholder categories).

- Participatory planning phase

4. Problem census (with each individual stakeholder group).
5. Compilation of problem rankings by facilitators (separating natural resource problems, combining stakeholder group rankings).
6. Plenary with stakeholder representatives and local leaders (to review problems, vote on the top three or four for solution analysis).
7. Solution and impact analysis (with each individual stakeholder group).
8. Plenary with stakeholder representatives and secondary stakeholders (to present the whole process, identify feasible solutions, discuss institutional arrangements proposed by separate groups and next step).

- Implementation phase

9. Develop and adapt community organizations and institutions for fishery/common pool resource management.
10. Community organization develops detailed plan to implement solutions agreed in stage eight
11. Review of plans by wider community and adjustments to plan (to mitigate or avoid any adverse impacts, for example).
12. Implementation of action/management plan (for example, physical works, application of rules, monitoring).
13. Institutionalization of management arrangements including local policy support.

At the heart of the process are stages four to eight that involve participatory workshops with separate stakeholder groups and combined plenary sessions. These stages have been the main focus of the action-research, as it is here that the substantive consensus is built. However, this should be seen as one important phase in a larger process. In the more general sense, action research addressing problems of the community has its focus on steps 9 through 13 where both institutional arrangements and fishery management actions are tested and evolve through the efforts of the community with advice and facilitation from non-governmental organizations (NGOs), government agents and researchers.

Participation issues

There is an inverse relationship between people's willingness to express their views frankly, and the number and diversity of people participating. Individuals tend to discuss issues more freely on an individual basis than in public. Some reasons why people may not contribute ideas to a public discussion are: they do not consider their ideas valuable; they do not want to upset the status quo; they want to avoid offending others; and it is not traditionally or culturally acceptable for them to speak in a public meeting (e.g. women and young people) when it is for others (e.g. male elders).

Building relations with a few key individuals can help obtain information of the real workings of society. This can be validated by triangulation with what other individuals say. However, it is a slow process and involves no explicit public consultation or planning objective.

An alternative is for people to express their ideas in a less judgmental forum where they feel comfortable, like with friends or with people of similar background. The drawback, however, is that these views are not aired in public and do not contribute to shared understanding and mutual learning, and so there is no change in the status quo.

The PAPD method takes into account these issues through a series of linked separation and aggregation steps that together can result in a balanced view. The separation steps are exercises undertaken by each stakeholder group. The aggregation steps are facilitated plenary sessions where all groups are represented.

These principles and the PAPD method can be adapted for participatory reviews and improvements in management as well as for initial planning. The process shown in the figure is consistent with the CBOs that result from PAPD designing management systems as part of the PAPD in sessions where government co-managers are actively involved with the primary stakeholders, these systems can include data sharing and reviewing.

FIGURE 15
Thirteen stages and three phases of the PAPD process

Source: Sultana and Thompson (2004).

5.2.2.1 Frame surveys

Frame surveys are used to generate important information required both for management planning purposes and for helping design data collection strategies or systems.

In particular, they are used to define the water bodies and areas that will be included in the plan or monitoring programmes and to describe the operating characteristics of a fishery or (co-)management unit including information relating to the environment, fishers, boats, gear, markets, fisher communities, and institutional arrangements etc (see Section 5.2.2). Because the types of data and information collected using frame survey approaches may be similar to those generated for SL analyses, opportunities may exist to integrate survey and appraisal activities where appropriate.

In the context of designing data collection strategies, the frame survey will provide important information relating to the infrastructure and characteristics of the fishery that can be used to guide the selection of data sources and tools, sampling strata, and information for raising sample estimates. At a minimum the frame survey should aim to include:

• The stocks or fishery being considered and the area under the jurisdiction of the LMI (Section 3.3.1).
• Information relating to the fishing units or FEUs including details of total numbers; gear types and technology employed, and distribution in relation to landing locations (Section 3.3.2).
• Diurnal fishing activity and landing patterns (quantities) by season and location (Section 3.3.2).
• Details of markets and their locations, (seasonal) prices; fish distribution routes, fish utilization, fish processing and marketing practices, fish trade, local consumption, number of processors or marketing units; supporting infrastructure (Section 3.3.8).

Additional information might include:

• Information on environments, habitats and potential catchment influences on the fishery or stock or locations critical to the life history of the stock or species (Section 3.3.1).
• Socio-economic categories of fishers (professional, subsistence etc), their sub-categories (e.g. women, children) and other stakeholders (fish traders, leaseholders etc) corresponding to or dependent upon different FEUs (Section 3.3.3)
• Management roles and responsibilities and institutional arrangements (Section 3.3.4).

Examples of typical data types collected for frame surveys are given in Section 6 of DFID Project Report ref. R7042 which can be downloaded at http://www.fmsp.org.uk/r7042.htm.

Frame surveys often draw upon information from numerous sources including directly from the LMI or resource users using Rapid Rural Appraisal (RRA) (Pido et al. 1996) and PRA techniques (Maine et al. (1996) and Annex D IFAD 2002, http://www.ifad.org/evaluation/guide/index.htm), and by direct observations, dedicated surveys or censuses, and other government departments or ministries.

Information is often usefully summarized in the form of thematic maps (Figure 16) or cognitive maps (see Berkes et al. 2001). Further guidance on frame surveys can be found in FAO (1997) Caddy and Bazigos (1985), Bazigos (1983) and Mahon and Rosenberg (1988). Meaden and Do Chi (1996) (see http://www.fao.org/DOCREP/003/W0615E/W0615E00.HTM) describe the application of GIS systems for creating thematic maps. De Graaf et al. (2003) illustrate the application of GIS for creating thematic maps for the fisheries of Lake Volta.

5.2.2.2 Recording the management plan

A checklist of the categories of information that might be included in each local management plan is provided in Section 3.3. Agreeing a common format will facilitate the coordination of individual management plans and activities (Section 3.3.10) and ensure that explanatory variables common among all LMIs or co-management units are available for comparative analysis purposes (Section 3.5.2). The plan will usually be presented in a report format, and should be made available to all stakeholders. Graphical or pictorial summaries of the plan, possibly based upon thematic frame survey maps (see above), may help to ensure understanding among stakeholders. Alternatively, it may be necessary to produce two versions of plan: a technical version that includes all the information in detail, and a less technical version that includes information in a form that all stakeholders can understand (see Berkes et al., 2001 and Box 18).

5.2.3 Stage 3 - Identify common data needs including specification

At this stage, each stakeholder group will identify their own information needs which will later be compiled to identify common data needs in Stage 5.

Once the stakeholders have a formulated or revised management plan that describes the environment, resources and fisheries defines local management objectives, strategies and actions to take, and outlines roles and responsibilities for its implementation (i.e. Category 2 information), it should now be possible for each key stakeholder group to identify a provisional list of data needs or interests in relation to the remaining information Categories 1, 3 and 4.

FIGURE 16
Example of a thematic map illustrating the location of important landing sites and numbers of fishing vessels by category in Antigua, Caribbean

Source: Mahon and Rosenberg (1988).

Stakeholders do not need to think of all the information needs at this stage. As plans change and as stakeholders develop their ideas and their understanding of the plan, the data and information needs will change also and can be updated. This means that the data collection system and management plan should evolve and improve together through time as they are revised or reviewed.

5.2.3.1 Identify data for Category 1 and 4 information

Guided by explicitly defined objectives, both government and each LMI should be able select a list of performance indicators and corresponding data types and variables for monitoring and evaluating policy and development plans (Category 1), and local management plan performance (Category 4). Sections 3.2 and 3.5 provide typical examples of what might be selected for monitoring.

Identifying indicators for each information category can be done by fishers, LMI members, local government officers and national fisheries officers (for example) working in separate groups to identify their information needs, or they can work in mixed groups and identify the information needs for all groups. This may be done through a workshop and may need to cover basic concepts.

To maximize acceptable data variable overlap, the stakeholder groups should be encouraged to identify several alternative indicators. Stakeholders should then indicate how important each indicator is to them by ranking or scoring the indicators according to their overall importance. Scoring can be done using a scale of 1 to 5.

A provisional list of explanatory variables corresponding to each performance indicator should also be identified at this point for inclusion in either routine monitoring programmes or to be fully documented and described in the management plan. This list should include variables to explain both the performance of local management plans and policy and development activities. Intermediaries may be well placed to help the main stakeholders identify these variables. Discussions and presentations based around ‘Bio-resource’ flow-diagrams and other conceptual models of the fishery, and the construction of a hypothesis matrix may help facilitate this process (Section 3.5.2).

5.2.3.2 Develop alternative indicators if required

It is unlikely that the indicators described in Chapter 3 will meet the needs of all stakeholders in every context. In these cases, stakeholders will need to select or develop alternative indicators that meet their requirements. Differences in the perception of appropriate indicators even at the community level (according to gender, age, occupation, wealth status etc) may also exist and therefore negotiation even at this level may be required. Alternatively several different indicators may need to be monitored to reconcile differences in perceptions or understanding of individuals.

The UNDP identify a number of desirable characteristics that any indicator should:

• be developed within an agreed upon conceptual and operational framework;
• be sensitive insofar as that a small change to be measured should result in a measured change in the indicator;
• be clearly and consistently defined so as to be un-ambiguous or lend themselves to various interpretations, or to give inconsistent results in different situations;
• be specific and measurable in that they have an explicit scale ranging from undesirable states to desirable states (along with specific weightings) that enables them to be used for assessment purposes;
• be policy oriented so as to provide practical information by being able to record either changes in the means recommended by policy or changes in the development impact attributable to policy;
• have ownership by users;
• reflect input, output process, and outcomes or impact; and
• be readily collectable and, thereby, lowering the technical and collection costs. Preference should be given to indicators for which existing data-collection mechanisms exist or can be adapted to fulfil the purpose of collecting data.

Further details can be found at http://www.undp.org under “Sustainable Livelihoods: Concepts, Principles and Approaches to Indicator Development”. Abbot and Guijt (1998, p. 41) suggest similar criteria.

The selection of relevant indicators for monitoring performance in relation to objectives can be guided by performance questions (IFAD, 2002, Section 5, downloadable at http://www.ifad.org/evaluation/guide/index.htm). Performance questions help focus the selection of appropriate indicators by addressing what is necessary to know if the management plan or policy is performing as planned. Once you have your performance questions, you can more easily decide what information you need to monitor. The following question can help find a good performance question: “What questions would you need to answer to know the extent to which you are achieving the objective?” After the performance questions are agreed, it is then much easier to decide what information you need to answer them.

Be sure to avoid duplication. Organizations such as national statistical bureaux, census bureaux, statistical offices or ministries of agriculture etc) may already be monitoring relevant indicators or data variables (see Section 5.2.4). Rai (1998) describes how indicators can be developed in a participatory manner on the basis of pictures depicting goals or management activities.

5.2.3.3 Identify the data needed

The next step is for the stakeholders to identify, for each quantitative (numerical) indicator, what data they would actually need to collect to be able to calculate the indicator. To do this, firstly identify the types of data required for each indicator. For example, to calculate indicators of fish abundance such as catch per unit effort (CPUE), the data types required will be catch and effort.

Next, groups should identify several possible data variables that could be collected for each data type. For example, catch data can be collected in several ways: as the weight of the catch, the number of fish caught, or the number of baskets or boxes of fish. Part I illustrates this process using a series of tables.

To maximize the opportunity for sharing data and the responsibility for collecting it, each group should attempt to identify several alternative acceptable data variables for each quantitative indicator. For qualitative (non-numerical) indicators and explanatory variables, stakeholders should discuss and negotiate as many common indicators and explanatory variables as possible. Grass roots indicators derived by individuals, households or communities (Abbot and Guijt, 1998) may not always be compatible with the needs of other stakeholders. Thus indicators will need to be selected or developed that integrate these different perspectives. This will involve discussing their significance and relevance with each stakeholder group. While a common set of indicators is desirable, bear in mind that different quantitative indicators selected by stakeholders may rely on the same data variables for their calculation. In these cases the emphasis should be on negotiating indicators that share common data variables.

Also bear in mind that the process of negotiating indicators requires developing a common understanding of success that is likely to differ among stakeholders. Negotiations of this type are beneficial since they can reinforce a shared vision for management and development. Interestingly, Abbot and Guijt (1998) report that stakeholders are often keen to work toward standard national procedures for monitoring and data handling. Note that even if local managers are not involved in data collection activities, selected indicators must be relevant and palatable.

5.2.3.4 Agree data specification

Having identified acceptable indicators and data variables for their calculation, the specification of each indicator or explanatory variable should be agreed among each stakeholder group and stated explicitly. This information is required to help design the data collection strategy (see Stage 5) including the selection of appropriate data sources, data collection tools, sampling units and sampling strata; and the identification of the required sampling intensity (sample size and sampling frequency) and coverage (sample or complete enumeration).

These specifications should aim to include details of the required frequency, accuracy and precision of the indicator or explanatory variable as well as details of any required standards (Sections 3.1.2. and 4.5). Some basic concepts related to accuracy and precision are presented in Box 20.

The required minimum level of accuracy (1- β) is typically 80–90 percent, but precision requirements will depend largely upon the how the indicators and variables are analysed and used. For example, catch data might need to be monitored with high precision (e.g. δ=10 percent) and accuracy to adequately evaluate the performance of different management strategies or stocking programmes (see Section 3.5.1.1). For policy and development planning purposes however, less precise (e.g. δ=20 percent) and less frequent estimates may be acceptable thereby providing opportunities to collect the data using less costly indirect methods and sources such as infrequent national censuses (see Section 4.6.1). Further guidance on data specification is available in Sections 3.1.2 and 4.5.

Stakeholders should be encouraged to clearly justify the reason for the data specification. This exercise can be very enlightening and may encourage managers to review or re-consider their roles and responsibilities as the purpose of management data and information becomes clear.

FAO have produced a number of very useful manuals and handbooks on the subject of sampling design including Stamatopolous (2002; 2004) and Sparre (2000) that should be consulted at this stage of the design process. These include detailed guidelines on how to achieve required levels of accuracy and precision, and on the design of data collection strategies, methods and forms. The FAO's ARTFISH software (see Section 5.2.7.2) also contains routines to help managers plan and design sample-based surveys.

5.2.3.5 Identify data and information to implement the management plan

Data and information required to implement the plan, particularly with respect to the enforcement of local rules and regulations and resolving conflict i.e. Category3 information (Section 3.4) might also be identified at this point, particularly if government is expected to take full or partial responsibility for these roles and responsibilities. Remember, most information required for helping to resolve conflict will already be contained in the management plan.

BOX 20 Accuracy and precision concepts
Accuracy (A) is a measure of how close the estimated value (m) is to the true value (μ) and can be expressed as:
where R denotes the population range (ymax-ymin).
Generally speaking, accuracy increases sharply with increasing sample size n expressed as a proportion of total population size N and then much slower beyond a certain critical sample size (equivalent to √N ) up to a maximum of 1 (or 100 percent). The figure below illustrates the form of this relationship based upon an arbitrarily selected population variance.
	For a given population variance estimate σ2, the minimum sample size to achieve a maximum allowable difference d between the estimated mean and the true value is given by
	where Z is the upper α/2 point of the standard normal distribution (approx equal to 2). Stamaltopolous (2004) provide estimates of n when σ2 is not known or cannot be estimated.
Precision
Precision refers to the closeness of a sample estimate to the expected value and, like accuracy, is a function of population variance, σ2. Estimates can be precise but not accurate when samples are not representative (biased). Precision is often expressed in terms of confidence limits around the estimate. Precision increases (the confidence interval becomes narrower) with decreasing population variance and increasing sample size. Precision determines the minimum differences detectable between means sampled between sites or periods of time. This Minimum detectable difference (δ) is therefore an important concept when evaluating the performance of different management strategies. The minimum detectable difference between two samples is a function of both the pooled population variance estimated by the pooled sample variance (S2p) and the sample size, n (Zar, 1984) given by:
where tα(2),v is the critical value of the t-distribution for a two-tailed test corresponding to the probability (α) of committing a Type I error (typically 5% or 0.05) with v degrees of freedom (2(n-1)), and where tβ(1),v is the critical value of the t-distribution for a one-tailed test corresponding to the probability (β) of committing a Type II error (typically 10 percent or 0.10) with v degrees of freedom.
Population variance will vary among different types of data variable. For example monthly CPUE estimated on the basis of catch per fisher will have a higher variance (be less precise) than catch per trap because individual fishers may own different numbers of traps and may not always fish the same numbers of days each month. A trade-off will therefore exist between the types of variables selected for monitoring and the required sample size. Larger samples sizes will be required for less precise data variables. Less precise variables may however be easier and less costly to collect. These trade-offs should be carefully weighed up when designing the data collection strategy (see below). Sampling the population prior to the design of the data collection strategy may provide useful estimates of the (pooled) population variance estimate σ2, i.e. the (pooled) sample variance, s2.

5.2.4 Stage 4 - Review existing data and identify shortfalls

Check that another institution or organization is not already collecting the data. Other government agencies, universities, research organizations, NGOs and other stakeholders will often have or be routinely collecting indicators or data variables of interest. It might be useful to start by asking whether reporting mechanisms at the village, district or national levels already exist for relevant information such as population, boat ownership, poverty indicators etc that may be generated by a range of methods such as national statistical or census or specific research methods. It is often helpful to compile an inventory, possible in tabular format, of existing information collection detailing what information is collected, who collects it, how, and for what purpose, and where the information is held (see Table 7 in Part 1 and Table 6-3 in IFAD, 2002).

Also determine how the information was collected, and whether it is reliable for the needs of each stakeholder. Remember, it may be possible to modify data gathering by other agencies to better support the needs of stakeholders (see Section 4.6.1).

5.2.5 Stage 5 - Agree data collection and sharing strategy

Co-managers should now be in a position to be able to begin designing the collection and sharing strategy to meet their data needs. This will involve identifying possible sources and methods for each data variable, identifying an appropriate sampling strategy and agreeing who will take responsibility for collecting the data and sharing it with whom.

5.2.5.1 Summaries needs and options

To facilitate this process, options to meet the requirements of each stakeholder identified above might to be considered in a tabular format similar to the type illustrated in below (Table 12). The table might initially include, for each stakeholder, details of:

• Required performance indicators or explanatory variables
• The required specification for each indicator or variable (frequency of estimate, acceptable accuracy and precision of estimate, any required standards, etc).
• The required data types and a list of acceptable data variables for collection. These will depend primarily upon the data type but also the operating characteristics of the fisheries determined by frame surveys or related methods and stakeholder capacity determined during the stakeholder analysis stage (Section 5.2.1).
• The population of interest (e.g. fishers in a village or management unit, households in a province or the entire fisheries sector.
• Survey coverage (sampling or complete enumeration) (Section 4.4)
• A list of potential sources (Section 4.1).
• A list of potential methods (Section 4.2). The selection of sources and methods should take account of available local capacity and resources, required accuracy and precision, but should also:
  • Be perceived by local participants as a way to help them address questions and problems.
  • Not impinge on participant's day-today activities and normal responsibilities.
  • Provide timely and necessary information for decision-making.
  • Produce reliable and, if not precise, credible results.
  • Reinforce community solidarity, cooperation and involvement.
  • Be gender sensitive.

5.2.5.2 Identify common data needs and agree responsibilities

Next, identify and highlight common data types or variables in the “Acceptable data variables” column of the table. In the example table below this has been done by highlighting common data variables with circles of the same colour.

For these common data needs, stakeholders should then discuss and agree who will collect that data, how and from where by selecting from the available options. Also agree with whom the data can be shared with. Consideration should be given to stakeholder capacity and their incentives to collect and share information (Section 5.2.5.5). In the table, open circles have been used to indicate who will collect the data, how and from where. The arrows connecting the coloured circles show how the data will be shared.

Having agreed a provisional data collection and sharing strategy, estimate the sample size and sampling frequency needed to meet the specified levels of accuracy and precision corresponding to each data variable. Another column should be added to the table to record this information. This may require pilot sampling programmes to find out the sample variance and if there is any sampling bias.

If stakeholders are unable to sample the population at the required intensity (sample size and frequency) to meet the common needs of the stakeholders then alternative sources and data collection methods may need to be selected. Sampling strata (Section 4.4.1) may be used to improve the precision of estimates, and may also be required for reporting or administrative purposes. Appropriate sampling strata to might be identified on the basis of the frame survey results. A further column may be added to the table to record any strata selected or required.

If proposing to use data that other institutions are already collecting, it will be necessary to negotiate the use of their data with them and agree how the stakeholders involved will access it. This will be particularly important if the data will be needed either more frequently or in a different format from that usually made available to the public.

Stakeholders will need to agree on the form in which the data or information will be shared. How will the collected data be compiled or analysed, and who will do what? This will depend on the level of detail required by each stakeholder and the capacity of each stakeholder to analyse and compile the data. Stakeholders should refer to the data specification for details of what is required by each group.

If necessary, in order to come to agreement on the data collection and sharing strategy, re-negotiate indicators, identify alternative data variables and adjust sample sizes until all stakeholder needs are met. Improvements to precision and accuracy may be achieved by selecting alternative data variables, and sources and methods. Where stakeholder needs in terms of accuracy and frequency do not coincide, each group will have to carry out their own data collection to satisfy their requirements.

A minimum ‘need-to-know’ approach may help make sure that the most important information is collected with enough accuracy and at the lowest possible cost. Once it is set up, the system can be expanded to include more detail on species, value, products and other factors (MRC, 2004).

If a mutually agreeable strategy cannot be identified, it may be necessary to create further incentives or re-negotiate the respective roles and responsibilities of each stakeholder. When considering alternative strategies, always bear in mind operational constraints such as the cost of salaries, training, capital costs (e.g. transport, computers, office equipment) establishing and maintaining information sharing networks (see Section 5.2.6) and any recurrent costs. Further advice on preparing budgets can be found in FAO (1999 p70); Sparre (2000, p159) and IFAD (2002 p7–36) that lists more that 40 potential cost items. Particularly for catch data, the selection of tools and sources may be guided by the use of fish disposition pathway diagrams that draw upon information gathered during the frame survey or PRA activities (see Section 4.6.2).

5.2.5.3 Uncommon (unique) data needs

Once stakeholders have agreed how to collect and share their common data needs, they should consider who might collect the remaining data needs. It may be that the government and LMI are happy to collect their own remaining data needs and can select appropriate sources and tools accordingly using the approaches described above. Alternatively, incentives may be offered to stakeholders, particularly the LMI or resource users, in exchange for the task of collecting data. Examples of incentives that governments are likely to be able to offer LMIs in return for collecting and sharing data and guidance on encouraging participation and data sharing are described below.

TABLE 12
Example of a tabular format to help identify common needs among key stakeholders and possible data collection and sharing strategies

Arrows indicate how variables might be shared to meet the requirements of other stakeholders. S-Sampling; CE-Complete enumeration.

5.2.5.4 An example

In the simple example illustrated in Table 12 below, an LMI has identified an interest in monitoring both CPUE and profit and has identified, with the help of other stakeholders, corresponding levels of precision for each indicator expressed in terms of minimum detectable difference (MDD) (see Box 20). It has also identified data variables that would be acceptable or collectable given its capacity and requirements, possible data collection sources and methods and corresponding sample sizes to meet its data specifications. The Provisional Fisheries Management Institution (PFMI) and the DoF identify poverty, employment and GVP as important indicators for policy and development planning and evaluation purposes. They also specify their own required levels of precision for these indicators, shortlist possible data variables, identified possible data sources and methods and estimated required sample sizes.

During a process of consultation and negotiation, the LMI agree to monitor catch weights and numbers of active fishermen in each month by direct observation and agree to share these data with both the PFMI and the DoF who can use these data variables to help determine levels of employment and calculate GVP. Sample sizes that can be collected by LMI were found to be adequate the meet the needs of both stakeholders. If this was found not to be the case, then it may have been necessary to negotiate larger sample sizes with the LMI or encourage the LMI to collect more precise measures of effort such as total fishing hours. In return, the PFMI agrees to share income data with the LMI that they plan to collect by means of interview methods on a monthly basis from households including those belonging to members of the LMI. The DoF offer incentives of the type described in Section 5.2.5.5. The outstanding data needs to construct the poverty and GVP indicators required by the PFMI and the DoF (i.e. living costs and price data respectively) have been identified as being available from other government agencies (Section 5.2.4).

5.2.5.5 Stakeholder capacity and incentives to collect and share

Local managers have an incentive to collect data to evaluate the performance of their own management plans allowing them to see for themselves the benefits or impacts of their management activities. Having been involved in data collection activities, fishers and other local stakeholders are more likely to believe the results of any evaluation. Indeed, the research activities described in Preparation of this document in Part1revealed that local managers have an interest in monitoring their management plans in a more objective and systematic manner. Local managers should not, however, be expected to collect data, such as that required for policy and development planning purposes (Category 1 information), simply to relieve the workload of the fisheries department. Likewise, it would be unrealistic to expect the administrative levels of government to collect data on behalf of local managers solely for the purposes of local management plan evaluation.

In addition to providing any necessary training and capacity building programmes, a number of further incentives exist that could be offered or made explicit to the LMI by the administrative levels of government, in exchange for agreeing to participate in local monitoring programmes, maximizing data overlap and sharing data and information. These incentives may be financial but providing local managers with support to formulate, implement and evaluate their other local management plans may be more attractive, sustainable, as well as mutually beneficial:

• Helping local managers formulate their management plans. This might include:
  • o undertaking baseline studies, frame surveys and livelihood appraisals to help local managers identify appropriate management strategies or pursue alternative livelihoods (e.g. tourism), and design effective data collection strategies.
  • o Provision of technical advice or information including for example, best practice for stocking programmes, management strategies, fish diseases, fisheries law (including citizen's rights), credit policies and alternative markets for fish or sources of raw materials.

As well as encouraging local participation in monitoring programmes, involvement in local management planning also provides opportunities for government departments to ensure that local management objectives are consistent with policy goals or objectives, and that the content of the plan is compliant or consistent with national legislation (Section 2.3.1).

• Helping local managers implement their management plans. This might include:
  • Helping local mangers enforce local rules and regulations.
  • Providing conflict resolution mechanisms.
  • Coordinating local management plans to minimize conflicts and promote integrated approaches to management (Section 3.3.9).
• Helping local managers evaluate and refine their management plans. This might include:
  • o Facilitating communication and learning among LMIs in support of adaptive approaches to management plan performance evaluation (see Section 3.5.4).
  • o Training and capacity building to help local managers evaluate for themselves the impact or performance of their management activities.

As well as encouraging local participation in monitoring programmes, these activities may also help government evaluate their co-management policy (see Sections 3.2.2.11 and 3.5.8).

• Communicating the importance of local monitoring in shaping policy.
  • Emphasizing that policy and development planning decisions will ultimately shape their livelihoods may provide local managers with a strong incentive to participate in local monitoring programmes, thereby ensuring that policy decisions take full account of the value of their fisheries resources. Communicating these benefits will be key during the management planning phase (Section 5.2.2). Local management institutions in Tanga participate in local monitoring programmes largely for the purposes of lobbying local government and policy makers (Purvis, 2004).
• Feedback The strongest incentive to participate from the perspective of local stakeholders may be the ability to see that their action, input or voices have impacted on policy and higher level institutions. This reinforces the importance of regularly feedback of the impacts of monitoring programmes (Box 21).
• Other incentives. Participation in local monitoring programmes could be made a condition of access to the fishery or licence agreements as part of the local management plan. Alternatively, payments for access or licences may be used to fund monitoring programmes either undertaken by appropriate administrative levels of government or the LMI. Catch records may even be used as collateral for credit!

5.2.5.6 Disincentives to participation

As well as promoting the incentives described above, government should also recognize and attempt to minimize any disincentives faced by local fishers and other stakeholders to participate in local monitoring programmes. Typically, these disincentives will centre on the opportunity costs associated with their participation, often, though not exclusively in the form of lost earnings. For example, earnings may be lost by fishers because part of their day must now be spent monitoring and recording leaving less time for fishing and related activities such as maintaining their gear. Earnings may also be lost if the value of their catch is diminished as a result of participation. Delaying the transportation to markets or sale to fish traders may result in the deterioration of their catch and subsequent loss of value, or the attainment of less favourable prices. These types of opportunity costs are often overlooked by designers of participatory monitoring programmes.

For example, on Lake George in Uganda, the opportunity costs of participation compared to the perceived benefits may be threatening the sustainability of the Catch Assessment Survey (CAS) (Lamberts, 2004). With poor preservation facilities, fisherman experience a decline in the value of their catch as they queue at “weighing check-points” to have their catches sampled. In addition, the time forgone waiting could have been used for other income generating activities. Opportunity costs should also be considered when co-designing data and information sharing systems and other co-management activities.

5.2.5.7 Other conditions that affect participation

In addition to perceived benefits and costs, a number of other factors described by Garaway and Arthur (2002) and Guijt et al. (1998) are likely to affect stakeholder participation in monitoring and evaluation which should be borne in mind when designing systems:

• Relevance of the PME to the priorities of participating stakeholders. Local participation will only be sustainable if it contributes to local understanding and empowerment.
• Timeliness and relevance of feedback of findings.
• Flexibility of the PME process to deal with diverse and changing information needs.
• Meeting expectations that arise from PME, for example acting on recommendations that are made.
• Degree of trust between stakeholders.

5.2.5.8 Check the agreed data collection and sharing strategy

A checklist of the type below may help to confirm or otherwise that you have the right strategy.

• Feasibility. Do you have the capacity, motivation, skills and equipment? Can you cover the geographic area adequately? Can sufficient technical support and training be provided?
• Appropriateness. Do stakeholders agree that the strategy is appropriate and do they understand it? Can it be supported by existing institutions?
• Validity. Do stakeholders who are to use the information believe that the methods are valid and generate sufficiently accurate information?
• Relevance. Does the strategy generate the required information? Are all the data relevant/required? Is there a pilot phase to test and refine the system (see below).
• Sensitivity. Are minimum detectable differences in indicator or variable estimates adequate for management purposes? Can it be adapted to changing conditions without excessive loss of reliability?
• Cost-effectiveness. Do sufficient resources exist to support the strategy? Will it produce the required information at relatively low cost or do cheaper alternatives exist that would be adequate?
• Timeliness. Does the strategy generate data in time for its intended purpose or use?
• Sustainability. Will it be sustainable without continuous project support? Is the system documented (see below) so that everyone knows what it generates and what information is disseminated?

5.2.6 Stage 6 - Identify or develop data and information sharing system

Once stakeholders have identified their common and unique data needs to evaluate the performance of their policies and local management plans and agreed who will collect these data, how, from where, and share with whom, they will now need to design communication systems pathways or networks for sharing these common data needs. However, depending upon the agreed roles and responsibilities of the stakeholders, these information sharing pathways or networks might also be required to:

• share information contained within local management plans (Category 2 information) with relevant administrative levels of government to help them coordinate local management plans and resolve conflicts with local managers (see Section 3.3.10 and 5.2.3.5);
• share Category 3 information with relevant administrative levels of government if they are expected to take full or partial responsibility for enforcing local rules and regulations described in the management plan (see Sections 3.4.1 and 5.2.3.5);
• share Category 4 information with relevant administrative levels of government, intermediaries and research institutions if they are expected to help local managers evaluate the performance of their management plans (see Sections 2.3.1 and 3.5.4);
• facilitate communication and learning among LMIs (or management units) to help local managers evaluate and refine their management strategies and institutional arrangements described in their management plans (see Box 24 and Section 3.5.4);
• feedback the following information to relevant administrative levels of government and LMIs:x
  • details of policy and development plans and the results of policy and development plan performance evaluations;
  • information about the actions needed to coordinate and enforce local management plans;
  • technical and socio-economic advice about effective management strategies, interventions and institutional arrangements generated by research including comparisons of management performance indicators and explanatory variables among LMIs or management units (Section 3.5.4).

Remember, many of these information sharing and feedback requirements of the system may be incentives offered in return for participation in data collection (and other co-management) activities (see Section 5.2.5.5). An effective data sharing system may therefore be key to sustaining the data collection system.

Advice on designing information sharing systems in the context of fisheries management is sparse. This probably reflects the fact that there are no blueprint solutions or generically applicable networks. The design of the information sharing system will be dependent upon the institutional arrangements and administrative structures defining the co-management arrangements, and the roles, responsibilities and capacity of each key stakeholder. Garaway and Arthur (2004) suggest that stakeholders should start be examining existing communication networks; their opportunities and constraints. A diagram illustrating potential information flows is shown in Figure 17.

5.2.6.1 Communications mapping

One way of identifying possible information sharing networks is to draw a communications map showing existing and required information flows among stakeholder. Start by writing each stakeholder group on a piece of paper or card and positioning them on a larger sheet of paper. Then draw arrows that link different stakeholders to represent current and required information flows (see Figure 18).

These diagrams can be drawn by stakeholders and used as the basis for discussions to identify their opportunities and constraints, and develop networks for the data collection and sharing system. Trust among stakeholders is the most important thing to be able to develop effective systems (Box 25).

5.2.6.2 Develop information sharing systems

To design and agree on a data and information sharing system, stakeholders will need to discuss and agree on the following points:

• who will share what data and information (and in what format) with whom;
• how it will be shared, distributed or delivered to the relevant people and organizations (e.g. on what media - printed matter, on CD, by radio, by post, delivered by bicycle, called through on mobile phone etc.);
• how frequently it will be shared, e.g. daily, weekly, monthly, annually. For example, if you are holding meetings, choose a time and frequency when people will be able to come.

FIGURE 17
Illustration of the potential information flows that might be created

Modified from Garaway and Arthur (2004)

FIGURE 18
Example of a communications map

It is important that these systems are realistic and sustainable. For example, if information is to be passed on through a phone call, there must be resources available to pay for the cost of the calls; if the postal system is unreliable, other methods will have to be found to deliver data on printed material, diskette or CD.

Some examples of existing data and information sharing networks are described in Annex 2.

5.2.6.3 Media for information distribution

A variety of media and methods might be used to deliver, disseminate and facilitate information sharing each having associated advantages and disadvantages. Literature, for example, can reach the largest number of people, but does not provide the opportunity for feedback. Radio provides for speedy communication to a wide audience and often encourages feedback, but is not appropriate for communicating detailed or complex information (Muthiah, 1991). Face to face contact can lead to greater understanding and more frank discussion and feedback. A matrix based approach involving discussions with the key stakeholders may provide a useful means of agreeing which approach might be most appropriate (see for example Table 13 below). Further practical guidelines are given in Maine et al. (1996).

The utility of existing information sharing networks might also be explored. The STREAM initiative, for example, facilitates information and knowledge sharing on a regional scale concerning a wide range of issues related to aquatic resource management and poverty alleviation. Stakeholders in ten Asia Pacific countries currently share information about technologies, practices, ways of working, research, development, legislation and policy via the STREAM journal and Web site (http://www.streaminitiative.org/). OneFish (http://www.onefish.org/), an Internet portal ,also provides an enabling environment for sharing research-based knowledge among diverse stakeholder groups.

5.2.6.4 Practical considerations when disseminating information

The following recommendations have been identified from IFAD (2002) and from the studies described in Preparation of this document in Part 1:

• Ensure messages are clear, understandable and relevant to the target stakeholder. The interests and concerns of different stakeholders vary and will require appropriate media, language and content. The format should be simple and easy to understand. The format required locally may differ from that of the government.
• Agree on the frequency for communication information. For example, if meetings are employed, choose a time and frequency when people will be able to come.
• Ensure timeliness. Ensure information is presented in time for its intended purpose.
• Make effective use of graphic information to facilitate analysis (Maine et al., 1996). Visually presented information is often easier to understand. Photography or videos may be effective (Box 26) but can be more costly. When disseminating the outcome of evaluations (see above), it may be useful to provide target stakeholders with prepared data sets which they can analyse themselves and present findings to each other, instead of simply presenting results. Remember, involving the collectors directly in the analysis of data, and presenting information back to collectors as soon as possible as helps to create a sense of ownership in the data, builds capacity and gives people a stake in the process (Garaway and Arthur, 2002).
• Standardization is only required when information has to be compared over time or between cases.
• Costs should be low, and never exceed the benefit gained from utilizing this information.

TABLE 13
An example of a matrix approach to help identify appropriate communication media and methods with stakeholders (modified from MRAG, 2004)

TABLE 14
Example of a monitoring and evaluation matrix

Source: IFAD, 2002.

5.2.6.5 Record the plan

The operational details of the data collection and sharing strategy should be summarized and supported by a workplan or schedule for data collection, possibly in a similar format to the Monitoring and Evaluation Matrix below (Table 14), with maps or diagrams showing the position of data collection locations. Providing a list of data collectors and a roster may also be needed. Annex C of IFAD (2002) provides a worked example in the context of project monitoring and evaluation.

5.2.7 Stage 7 - Design data recording and management systems

5.2.7.1 Data recording systems

Data can be recorded in many ways, depending in large part on the data collection method (Section 4.2). Some methods, particularly the interview and direct observation methods may require filling in of logbooks, ledgers, forms or tables, whilst others might utilize a video camera or taking detailed notes. For each data variable selected above, it will be necessary to agree upon how it will be recorded. Consistent methodology will help to ensure that data are comparable. For routinely collected data, recording forms should include basic information that facilitates data checking and ensures that data can be referenced, sorted, collated and manipulated (Table 15).

TABLE 15
Examples of basic fields of information typically included in data collection forms

While it is impossible to develop generic data collection forms, Halls et al. (2000) (see report Ref R7042 at http://www.fmsp.org.uk/r7042.htm) identified fields of information that are typically included in recording forms that are employed for frame surveys, catch assessment surveys (CAS), biological and socio-economic monitoring programmes, and for vessel monitoring and licensing purposes. While these fields are not exhaustive or definitive, workers have reported that they provide a useful starting point for designing data collection forms (Jim Anderson pers. comms.). Users of data recording systems should ideally be involved in their design to ensure that they are both understandable and appropriate. Further guidance on the design of data collection forms including numerous examples can be found in Sparre (2000); FAO (1999); Caddy and Bazigos (1985) and Bazigos (1983).

5.2.7.2 Data management system

When considering how to store data and information IFAD (2002) recommend that stakeholders bear in mind:

• The data and information to be stored. In principle, everything that co-managers decide to monitor and evaluate will need to be stored in some way either as reference, for tracking change through time or for making comparisons among different locations or sites. Only store data and information if it will be used. Some raw material generated by PRAs may not need to be stored if the information generated from it has been processed and stored elsewhere. Diagrams for example, may not need to be copied, distributed and stored at all levels. Originals can be left with the stakeholders who produced them.
• The needs of different stakeholders. How the data is stored will depend largely upon who needs to have access to it and how frequently. Consider their capacity and the types of communication methods which they are comfortable with (see Section 5.2.6).
• The format for storage - hard copies or electronic data. Generally speaking, electronic files allows access to the data by more people. However, not all data or information gathered at the local level can be readily formatted electronically. Furthermore, local level stakeholders might not have the capacity to access computers or electronic networks. Information presented diagrammatically or generated through discussion may be distilled into short reports for storage and dissemination.
• The need to regularly review the content of the system. The content of the data storage system should be reviewed regularly to avoid it becoming congested and unwieldy. Computerized data should be regularly archived yet remain accessible. Documents may need to be stored for legal reasons or for accounting purposes. Copies or material required for making comparisons through space or time must also be retained including baseline data, copies of management plans, and summaries of progress, etc.

5.2.7.3 Electronic databases

Electronic databases provide a means of storing raw data in a secure and standard format and help facilitate its rapid processing (filtering, aggregating, transforming) for decision-making. Databases also help to ensure data validity, integrity and consistency, and may allow different datasets to be integrated thereby increasing their overall utility. FAO (1999) offer guidelines on database design including advice on software development, interfaces, documentation, as well as data processing, reporting, access and dissemination. A more thorough treatment of the design of database tables, forms and queries can be found in Sparre (2000) which includes a detailed description of an example database.

FAO have also produced software called ARTFISH (Approaches, Rules and Techniques for Fisheries Statistical Monitoring) to help design shore-based surveys for generating fishery production and value information, and supporting database (Box29). The software can be downloaded at http://www.fao.org/fi/artfish.htm which also contains links to other relevant documents.

5.2.8 Stage 8 - Implement, evaluate and refine

The final stage in the design process is the implementation and revision or refinement of the system during which consideration should be given to the following:

• Training and capacity building. It is likely that considerable attention and resources will need to be directed towards training and capacity building for staff and other stakeholders to ensure successful implementation. For example, if local stakeholders are to be involved in selecting or negotiating indicators, then training on their meaning, selection and use may be required. Similarly, training on species identification may also be required. Training and capacity building needs should be assessed against each indicator and corresponding set of sources and methods (effectively each row in Table 12). The gaps identified should form the basis of a training plan. Guidance on the scope and delivery of appropriate training programmes including advice on “training of trainers” are available in FAO (1999) and Sparre (2000). As well as providing training, it will also be necessary to ensure that everyone involved in the system has sufficient financial resources and equipment to support their activities.
• Incentives. In addition to the incentives used to encourage participation by local stakeholders described in Section 5.2.5.5, attention should also be given to ensuring that all staff involved in the data collection and sharing system are motivated to participate effectively. Incentives to encourage participation and motivation include:
  • Ensuring that all stakeholders understand their respective roles and responsibilities, supported where appropriate, by memoranda of understanding or contracts.
  • Regularly demonstrating the function and value of the data generated including the results of evaluations, and explanations as to how the data are analysed and for what purpose (feedback). Local stakeholders are often reluctant to disclose information concerning their fishing activities and earnings because of fear that statistics will be used for taxation purposes.
  • Financial and other rewards such as housing and vehicle use.
  • Activity support: provision of adequate finances and logistics to conduct data collection and sharing activities such as transport, accommodation, species identification manuals, raincoats, weighing scales, maps and rosters.
  • Recognition: listening to staff and participants and acting upon their recommendations, possibly on the basis of staff appraisals.
• Technical Committees and Legal Frameworks. Prior to implementation, FAO (1999) also recommend that technical committees be established to guide the development and refinement of the system. Furthermore, if the provision of data is to be made a condition of access or licensing agreements (see Section 5.2.5.5) then legal or policy instruments need to be in place before the data collection system can be implemented.
• Data Verification. Fisheries data are prone to error and therefore data verification is necessary to ensure that data generated by the system are both accurate and complete. FAO (1999) describe methods for verification according to data type and the scale over which estimates are made. “Triangulation” where at least three sources or techniques must be consulted or used to investigate the same topics is a common approach employed to verify the accuracy of data or information generated by RRAs and PRAs methods (Pido et al., 1996). IFAD (2002, Section 6.3) provide useful guidance on how to improve the reliability of data and information including tips to avoid non-sampling errors and methods to verify data.
• Feedback. The data sharing component of the system should have already addressed the feedback requirements of the different stakeholders (Section 5.2.6). However, feedback should also be encouraged to help identify inadequacies or weaknesses in the data collection and sharing system which can then be addressed immediately and monitored. The MRC (2004) suggest that the results of information gathering programmes should always be fed back to those involved in the data collection activities providing opportunities to discuss and revise any shortcomings, inaccuracies and inconsistencies of information and data, which otherwise would become untrustworthy.
  • Pilot data collection programmes. Where possible the data collection and sharing strategy should be pre-tested on a pilot scale. Sparre (2000) describes a 42-step process for piloting a data collection programme which includes consultation, training, resource need identification, design, data verification and feedback activities.
  • System Appraisal. The system should be reviewed continuously in a participatory manner to ensure that it meets the needs of all the stakeholders involved in the co-management process. Reviews should include a continuous process of data verification (see above). Sparre (2000) describes a detailed iterative process to system design appraisal and refinement that FAO (1999) suggest should “…give a higher probability of system adequacy and stability”.
  • System Documentation. Documenting the system is important to keep existing and future stakeholders informed of the activities and status of the data collection and sharing system, to justify further investment in the system, and help to make comparisons with other systems. Sparre (2000) lists 13 documents that might be relevant including guidance on their format and publication. Reyntjens-Mesquita and Pittaluga (2004) recommend that documentation should at least include the following headings: Objectives; Indicators; Sources of information; Baseline data needed; Who is involved?; Tools and methods; How often need [Data collection frequency]; How often will the data be used, analysed, and by whom; Who receives the information.