3.4 Participatory action research: basic methods and techniques

Participatory action research includes an increasing array of techniques and tools designed for dealing with a variety of management issues and problems. For simplicity of presentation, all of these techniques can be brought back to three basic methods for generating information and making decisions. These three methods, which have been widely used and tested by applied social science, are: participant observation, individual interviewing, and facilitated group meetings and exercises.

Participant observation

Participant observation consists of taking part in social situations with the aim of discovering issues, events and interactions which may be obvious for the local actors but unknown to the external observer. Extensively used by anthropologists, participant observation has proved to be especially useful in the framework of participatory action research for:

• understanding the way in which local people use natural resources and manage their environment;
• understanding the local social environment, with special emphasis on status, roles and behaviors related to gender, age, wealth, social status and ethnic differentials;
• understanding the daily relationships and communication flows existing among non-local professionals (e.g., project staff) and different groups of local people;
• analyzing daily, weekly, monthly or yearly patterns of activities and the time allocated by local actors to perform them.

As used in conventional anthropology, the technique of participant observation typically entails a regular routine of being there - taking notes, reflecting, hypothesizing and making many repeated observations for confirmation. Such a time-consuming procedure does not usually fit well the practical needs of participatory action research. While failing to provide the in-depth understanding of local society and culture which comes after several months of full-time participant observation, a few days spent walking around and visiting people and places with an open mind are a good entry point for a professional new to a particular community or project setting. The discussion of insider/outsider experiences in the participant observation method is a key component of any action research process (see Box 3.6).

To get the most out of a rapid participant observation of a local setting, PAR practitioners tend to focus this practice on situations directly relevant to specific issues of interest. For instance:

• transect observational walks (see Annex B, section B.1): walking with one or more residents across the area surrounding a village can be useful for obtaining information about patterns of natural resources and their exploitation or use;

• visiting places where different community groups meet: people tend to gather in places like markets, coffee shops, football grounds, religious buildings, water source points (boreholes, springs), etc. Observations can be made at such locations to identify patterns of social interaction among different groups;

• attending official meetings and ceremonies (if possible and acceptable): observing at events, such as community meetings and ceremonies can generate insights about the political and social dynamics in the community;

• observing service delivery and project activities: many kinds of services are provided in communities; observing the process can help in assessing the reciprocal attitudes and behaviors of providers and users. Traveling along during regular project activities, such as agricultural extension visits, natural reserve patrolling or family planning outreaches is a very good way to understand the communication flows between external agencies and local people. While it is desirable to remain neutral relative to either the project or the community, assisting with some simple tasks during the activity is generally appreciated by staff and recipients. In this way, participant observation can lead to closer rapport and good opportunities to engage in discussions with local people, service providers and project staff.

In most observation sessions, note-taking is appropriate and necessary. Jotting down what has been observed is not only an aid for memory but also a way to focus the observer's attention on issues directly relevant to the problem(s) for which the action research is being done. Pre-set observational check-lists can be used, but they are more controversial. The advantages - standardization and systematization of observing specific items - will need to be weighed against the risks of losing an open-ended, discovery-oriented and relaxed approach which is a basic strength of participant observation.

As mentioned earlier, participant observation is usually combined in action research with various individual interviewing and group discussion techniques. Indeed, verbal interaction on the spot, i.e., talking with the people who happen to be there, can help to elicit insider interpretations of an observed behavior, event or situation. Most importantly, as PAR is principally for the benefit of local people, any observation collected by outsiders needs to be communicated to the insiders and discussed/evaluated together with them.

Interviewing - for qualitative and quantitative information

Interviewing is basically a process of inquiring into another person's perceptions about a particular issue (or set of topics). It can include exploring their knowledge, feelings, attitudes, opinions, past experiences and expectations for the future.

Interviewing can be useful for a wide range of concerns, such as:

• local knowledge of the natural and social environment: natural resources available and their traditional use; local technologies; means to mobilize labor; social organization of the community; gender and age issues; etc.;
• felt problems and needs: related to household economy, conservation of the resource base, education and health; access to reproductive health, family planning, credit services, etc.;
• perception of existing initiatives in the area (e.g., conservation or development projects): relationships among local actors and staff; relevance of the initiative to local needs; functioning, acceptability and appropriateness of the services provided; etc.;
• time trends in the life of the community: effects of seasonality on social life; distribution of activities throughout the year; local history; life histories; expectations about the future; environmental and population coping strategies in the past; etc.

Two main interview approaches are used by social researchers: open-ended and closed-ended interviewing. Open-ended interviews are based on a more or less organized and standardized sequence of questions whose answer are richer than a plain 'yes' or 'no' statement, a numerical figure (e.g., 15 years) or a categorical judgment (e.g., 'good' or 'bad'). As with participant observation, open-ended questions aim at discovering elements of the insiders' perceptions of the topic under investigation, which are likely to be almost unknown to the interviewer. Open-ended interviewing is thus a powerful means to catch 'qualitative information'. Widely used in anthropology, psychology and other traditions of applied qualitative research, open-ended interviewing is a key component of participatory action research (see Annex B, section B.6).

Choosing the proper respondents is an important element of open-ended interviewing. Some respondents could be selected because they have special knowledge about the topic of the action research - they are key informants. It also helps if they are ready to talk and they are somewhat reflective or analytical in their thinking. Other respondents might be selected for their representativeness as members of different interest groups, or on the basis of any special social status accorded to them, either formally or informally. Open-ended questions can also be addressed to a group of respondents. Groups may be spontaneously set up (e.g., a natural group of peasant women encountered while they are at a spring for water) or convened by the interviewer (e.g., a focused discussion group or focus group).

Most often, the sampling of respondents for open-ended interviews is based on judgmental (purposeful) criteria rather than random chance. Such purposeful sampling is quite different from the random sampling useful to draw statistical probabilities in quantitative research. Statis-tical sampling is based on randomness so that researchers can confidently generalize results from a small sample to a larger population. The power of purposeful sampling lies in selecting information-rich cases for in-depth analysis related to the central issues being studied. A list of different sampling strategies that can be used in qualitative action research (either for individuals or groups) is presented in Box 3.7.

The success of an open-ended interview depends very much on the communication skills of the interviewer, which are needed to keep the discussion as a relaxed dialogue. These skills include helping the interviewee (respondent, participant) feel at ease, phrasing questions in clear but not leading ways (i.e., not suggesting the 'correct' or 'expected' answer, as in the question 'Do you believe conserving natural resources is good for your community?'), and introducing probing questions appropriately (e.g., asking for further details when respondents give general answers).

Confidentiality can be an issue for some participants. Although the potential topics for interviews mentioned earlier in this section are not generally very sensitive for most individuals, the nature of the interview process, especially qualitative open-ended interviewing, can result in respondents disclosing information which they would prefer to keep confidential (i.e., they want the information revealed but not the source of the information). It will be important for the interviewer to respect such wishes if they arise.

Open-ended interviewing techniques are best carried out with detailed note-taking aimed at catching the exact words and phrasing of respondents' answers (i.e., quotes, verbatim statements). Tape recording can be of great assistance in this effort, but only where and when it is fully acceptable to the respondent.

Once data are collected, good summarizing skills and insight capabilities are needed to extract meaningful content from the statements provided by respondents, which at times can be long and wandering. The latter task can be facilitated by the use of qualitative analysis matrices (see Table 4.7).

The qualitative information provided by open-ended interviews of individuals or groups is often counterchecked and completed by other methods and techniques, including participant observation and closed-ended interviewing. This is especially important when it is desirable to assess the generalizability of the data for a larger group of people than the small numbers directly interviewed.

Closed-ended interviews use carefully organized questions which allow only a limited range of answers, e.g., a yes/no answer, a categorical answer (male or female, peasant or herder, etc.), a preference answer ('A is better than B', etc.) or an answer expressed by a number (age, date, time, distance, land measurements, etc.). The series of questions in this type of interview are usually presented with structured questionnaires. These are pre-set formats in which the interviewer marks responses against a given group of possible answers for each question. To allow generalizing from these kinds of results, different statistically random sampling procedures are at times used to select the respondents, with the aim of interviewing a representative sample of the population under study.

In applied social sciences, such as demography and epidemiology, structured questionnaires are widely used for collecting quantitative data (i.e., data which can be processed by means of arithmetic and statistical formulas). Many research practitioners share a critical view of rural surveys which are solely dependent on questionnaires (see Box 3.8).

Two major complaints about questionnaires relate to the loss of human touch and the extent of technical expertise required. Structured questionnaires are often poorly understood by respondents and there are problems with the reliability of the information collected. Moreover, carrying out high-quality quantitative surveys requires specialized skills for questionnaire design, as well as the capacity to cope with the sophistication and abstractions of statistical analysis.

PAR, however, may also need quantitative information (such as the data provided by a population census or a demographic survey). In other words, it may require a blend of qualitative data about the nature of a problem or a situation combined with quantitative data about the distribution or extent of the problem to ensure a full and reliable analysis of a given situation.

To meet these needs, PAR can indeed combine qualitative and quantitative methods. This can be done sequentially, e.g., by doing open-ended interviews first to assess the range and nature of responses, followed by closed-ended questionnaires to check on the prevalence and distribution of responses. Semi-quantitative and participatory methods of ranking (see Annex B, section B.8) can then help individuals and groups of respondents to express judgments and opinions about the results of a questionnaire. Discussions of specific findings can also be held with groups which are representative of specific social actors, interests and/or the community at large. These discussions can focus on the interpretation and validity of the findings derived from questionnaires.

Facilitating group meetings and exercises

Participant observation and individual interviewing are powerful ways to elicit information about a local social setting. The action research approach, however, does not stop at this point. As soon as possible, the 'research' interaction needs to become a partnership among local and non-local actors (community groups, local institutions, project staff, etc.) aimed at taking action - if action is due - on the basis of the research findings. The basic method which enables this to happen is facilitated group work (see Annex B, section B.14).

Participants for action research group exercises can be recruited through purposeful sampling strategies, such as those employed for open-ended interviewing. Another strategy is to use spontaneous clusters of community members on the basis of a common interest, e.g., women farmers, herd-boys, water vendors, traditional birth attendants, etc. Such natural groups - which often coincide with, or can lead to identify, specific interest groups - are actually the core actors of the PAR process.

Frequently, members of interest groups will already be participating in the organized or informal indigenous social groups that are typically found in most rural communities, e.g., farmers' groups, burial societies, self-help groups, mothers' unions, revolving credit clubs, etc. A core concern of many of these social groups is mutual aid or assistance. At times they constitute themselves as community-based organizations to support the socio-economic and environmental interests of their individual members or of the community as a whole, e.g., as a user group, a local cooperative, a village council, a residents' association, etc. Where such groups exist, they are very important actors to involve in assessing the local environmental and social situation and in planning some initiatives to ameliorate it (see Box 3.9).

Action research practitioners rely on a large selection of techniques and tools to promote participatory appraisal, planning, monitoring, evaluation and re-planning among the members of interest groups. They include group exercises that focus on:

• collection and organization of information owned by different local and non-local actors;
• discussion and validation of information collected in the field or from existing secondary sources (e.g., census, official statistics, technical research);
• building consensus, resolving conflicts, setting priorities and making decisions;
• learning by doing during the implementation of decided activities.

Experience has shown that group work with local actors is more likely to be effective and efficient when a neutral and coherent facilitation is provided by an experienced person who is not a member of the community. A good facilitator helps the participants to think and communicate with one another, and avoids imposing his/her personal view or suggesting the 'right' answer or decision either explicitly or implicitly. He/she takes care of helping the group achieve the results the group itself wishes to achieve, managing interpersonal dynamics, ensuring that the discussion is kept on a relevant track, and keeping the length of the exercise within reasonable limits.

To carry out this role, the facilitator needs a good understanding of the social and cultural reality of the participants. This awareness is essential so that his/her personal style can be adapted to the rules of behavior, the communication and analytical skills and the cultural attitudes of the participants. A good facilitator will be able to strike a balance between the need to achieve some positive results in the group work and the need to keep the interaction among participants as smooth and relaxed as possible. Maximizing the contribution of each participant in the group exercise, settling conflicts and building consensus are the facilitator's main responsibilities (see additional details about facilitation in Annex B, section B.14, and in Chapter 5).

As part of skillful and non-intrusive facilitation, the creative use of visual aids is an important strategy for supporting group exercises in action research. Some examples of visual techniques include the following:

• Maps and transect representations can be used very effectively in groups to describe and analyze the community's spatial distribution of features of special interest (e.g., natural resources, types of soil, vulnerable families, types of services, water points, land tenure patterns, etc.).

• Drawings, posters, pictures and slides as well as open-ended stories, popular theater and community-directed videos can be an excellent entry point for group discussions.

• Sorting, counting and ranking exercises may be done in written form but, if literacy is low, they can equally be carried out with everyday objects, such as seeds, stones or simple sketches on small slips of paper.

• Graphic representations by means of pie charts or bar charts (or

  better yet pictograms, i.e., graphs built of pictures) are suitable for conveying quantitative information even to non-literate participants. The pictograms (whose shape is often inspired by daily objects such as trees, animals, pottery or food) can be used to describe and analyze time trends, patterns of relationship among different actors, or sequences of causes, problems and solutions.

• Analytical tools (e.g., matrices, problem-cause-effect trees, Venn/ chapati diagrams - see Chapters 4 and 5) can be used to organize and analyze findings, including qualitative statements. They can also be used on flipcharts or chalkboards for assembling the ideas developed in a brainstorming session with a group.

3.5 Participatory action research: conditions for success and problems of validity

Participatory action research can provide effective support to any conservation or development initiative. Though based principally on common sense, PAR exercises have some pre-conditions, and entail several potential technical and attitudinal difficulties that are addressed in this section.

Community acceptance

An underlying assumption of the PAR approach is that the people in the community are willing and can afford to invest energy and time in the process. Another assumption is that the non-local partners and facilitators in PAR have the trust and confidence of the local community. This kind of trust is not developed overnight, and in some cultures it develops only over months, and perhaps years, of working together.

It is unrealistic for people unfamiliar with a community or the local context to have the kind of understanding and local acceptance necessary to become partners or facilitators in the PAR process. Therefore, it would be advisable for at least one or two of the non-local support team to take the time to get to know the community and build rapport with community members. Generally, this will mean either living within a community for a period of time, or maintaining contact with the community through regular visits. This contact could be on the basis of regularly scheduled meetings (e.g., those carried out by health workers or agricultural extension agents) or on the basis of a long-term conservation or development initiative that require frequent visits to the community for coordination. Once trust is established, it is far easier to collaborate with community members in PAR or any other activity.

To enable those non-local partners who have not had extensive contact with the community to get to know community members and vice versa, it may be useful at the outset of a PAR to engage in some kind of ice-breaking activity. For example, the non-locals might participate in traditional village tasks such as gardening, construction or clothes washing. This can serve to build rapport while at the same time helping community members to see that they possess a unique body of knowledge and skills not necessarily shared by outsiders. It may also help the non-local partners to address their work with more humility.

Staff technical know-how and attitudes

After taking the initial decision to adopt participatory action research (e.g., at the start of an initiative or to re-orient an existing project), some specific know-how can be transferred to its staff via training by an experienced practitioner. Regional and sectoral networks of concerned institutions may provide technical assistance and open fora for discussion and exchange of experiences. Learning materials in different languages - this manual being but one example - are also available (see the References section for recommended reading on PAR and PEC).

The availability of qualified support and literature alone, however, cannot ensure the deeper attitudinal change that the practice of action research demands from project managers, conservation and development professionals and field workers. While this change usually occurs spontaneously in the process of field experience in participatory action research, it can be facilitated by attention to a number of general cognitive and behavioral points:

• Knowledge is for the community, not for the non-local partners. The most important objective of action research is to strengthen local abilities to seek, organize and utilize relevant information to solve problems. Participatory exercises are not the time or place for academically oriented research. Collecting data in a participatory way may be initially greeted with interest, but later resented by the community as a waste of time if the results are not returned promptly and applied within their territory.

• Non-local partners in participatory action research need to have something to offer. To be useful and justify their presence, the non-local partners and facilitators in an action research process need to master their technical background as well as research and communication skills. Unskilled facilitators are useless or even detrimental and expose everyone to frustrating experiences. The best way to learn how to facilitate action research exercises is by doing it with an experienced colleague.

• Local culture deserves respect. An action research process which is conducted and managed according to local customs is much more likely to be successful in the eyes of all participants. Information should be handled carefully. Whenever requested, anonymity and discretion in sharing information with others should be assured. Communities appreciate respect for local traditions, e.g., following local protocols for introductions. At the same time, partners and facilitators should be aware that cultures and communities are dynamic entities: they change, and often rapidly. Attempts to restore customs and practices which the community has already dismissed are un- likely to succeed and may cause resentment. People do not like to be told what to maintain and what to change in their culture.

• Communication is a major concern. Conducting action research exercises in the local language (i.e., the variant of the national language spoken in the area or the local ethnic vernacular language) ensures full participation by the local community, regardless of education. Consecutive translation is tedious, boring to participants and not appropriate. If the PAR partners are not comfortable with the local language, bilingual facilitators can be recruited and trained.

As already pointed out, simple written materials (such as flipcharts) are often very useful, yet their value is limited in situations where literacy is low. Visual aids, such as drawings, maps, photographs, slides and videos, are always recommended, and they are essential when the majority of participants are non-literate. It is, however, a good strategy to test the cultural acceptability of these tools beforehand. At times, specific colors or images may be culturally sensitive or linked inappropriately with local political, religious or ethnic divisiveness.

As crucial as it is to recognize the central role of the local actors through these behavioral rules, participants and facilitators will need to avoid an oversimplified conclusion that local perceptions, knowledge and experience are the final or only truth. Just like academic institutions and development agencies, local communities are influenced by many forces that can bias their capacity to analyze their situation and make effective decisions for improving it. Prejudices, conflicts, corruption, privileges, resistance to change, indifference and discrimination are widespread among all types of societies, agencies, institutions and communities. Therefore, a good action research process engenders among its participants an open and possibly critical attitude towards both indi-genous and external points of view.

In practice, this means that an action research process entails an exchange and possibly a mediation between what local people think and wish to do and what is suggested by other sources of knowledge, including development professionals and various kinds of research. Where population dynamics and natural resource management are concerned, the findings of biological, ecological, medical, demographic, economic, social and cultural studies carried out by outsiders have an important complementary role together with local knowledge of the environment and society. Participatory action research methods can also be used in facilitating the presentation of technical findings to non-specialized audiences.

Validity and reliability of participatory action research results

In conventional research, validity is taken to mean how close the findings are to reality; and reliability is equated with constancy of findings. When it comes to participatory action research, the concepts are interpreted somewhat differently. In striving for sustainable development, PAR results may be considered valid and reliable when their application is environmentally beneficial and/or brings about an actual improvement of the living conditions of the people, which can be sustained over time with minimal cost to the environment. In other words, 'valid' and 'reliable' are understood from the perspective of local people. The results of the research have to be first of all meaningful and positive for them.

By starting from local knowledge and wishing above all to empower people to define what constitutes problems, opportunities and solutions for them, participatory approaches challenge the conventional tendencies to rely on 'scientific' knowledge and external authorities. Yet, validating findings is an important concern which, in PAR, is dealt with by a method known as triangulation. In a strict sense, to triangulate means to utilize at least three different points of view to analyze a given event or situation. More generally, triangulation is based on the idea that using multiple sources and methods is the best assurance of the validity, reliability and completeness of the collected information. Two basic modes of triangulation are used in action research: external and internal triangulation.

'External triangulation' involves a comparison between the information generated by participatory action research and data from external sources, such as censuses, official statistics, aerial photographs, or local independent research and technical studies. External triangulation is often based on a review of secondary data, i.e., information already existing and available from national and local agencies and academic institutions or published in papers and books. Less frequently, additional studies, such as quantitative surveys, are used to validate qualitative action research findings.

'Internal triangulation' refers to a set of techniques for strengthening validity within the action research exercise itself. These techniques include:

• Comparison of different perspectives. Different interest groups may have views and perceptions of 'reality' that are different from each other. One way to understand these co-existing views is for different interest groups to go through the same exercises, e.g., to compile their own separate lists of concerns and then - possibly - compare and discuss them (see Chapter 5).

• Use of different methods and techniques for exploring the same topic. For instance, a description of the way in which the community uses its territory may be developed through a combination of observational walks, interviews with groups and a participatory mapping exercise with community members.

• Involvement of non-local professionals with different disciplinary backgrounds. Different disciplines will raise different questions about the same issues, thereby stimulating deeper analysis by the participants. For instance, family planning needs among the women of a community might be assessed through a combination of open-ended interviews by an anthropologist, reproductive life histories collected by a nurse and a group discussion with the local women's association, where the first two types of information can be presented, reviewed and additional ideas elicited.

This discussion of external and internal triangulation helps to show that participatory action research describes and analyses problems or situations through an overlapping variety of techniques, perspectives and social interests. Therefore, multidisciplinary teamwork, capability to combine local understanding with scientific explanation, and readiness to shift from theoretical reasoning to concrete decision-making are key elements for the successful application of participatory action research in real-life settings.

Even when all these conditions are present, however, PAR practitioners will continually be challenged by difficult questions, such as:

• Is the local classification of soils, as explained by one well-experienced village elder, compatible with the optimal use of this resource?
• Can we assume that the traditional means of contraception in the region are safe and effective?
• To what extent do existing local institutions provide a structure for equitable and effective decision-making concerning the local environ-ment and population dynamics?
• How genuine is the consensus achieved at the last participatory planning or evaluation meeting?

It may be helpful to remember that the answers to such questions need to make sense for both the PAR facilitators and the local people, and that both questions and answers may be reframed and evolve with time.