Preparations
Interviewing
Probing
Triangulation
Direct Observation
Analytical Activities, Exercises and Games
Record Keeping
Analysis
Participatory Epidemiology and Womens Farming Activities
One of the most common epidemiological applications of PRA methodology is to gain a rapid overview of the range a communitys animal health problems and a measure of the importance that people place on each of them. This is achieved by directly asking livestock owners what animal health problems are occurring or have occurred in their livestock. They usually respond by naming a number of disease syndromes in their own language. The epidemiologist or appraisal team will ask the respondents to describe each syndrome individually and build a definition of each term. Once the description has been received, clarifying questions can be asked. After a number of interviews have been completed, consensus definitions of important animal health syndromes will have been determined. This is a relatively straightforward process, but the actual execution requires good interpersonal skills and an open mind. Further, the quality of results can be greatly affected by interviewing method and enhanced by utilisation of a few analytical exercises or activities.
This section will summarise the basic tools of participatory rural appraisal and suggest methods for their application to the completion of baseline studies of community animal health. A number of detailed documents and training courses are available on PRA and RRA in general (McCracken, 1988; IIED, 1994; Waters-Bayer and Bayer, 1994 and Narayan, 1996). The reader is encouraged to consult such texts for further information. The techniques that directly relate to animal health and have been tested in the field in epidemiological exercises will be discussed here. Many more techniques or variations of techniques exist and the potential for creative experimentation is only limited by the time and interest of PRA practitioners.
The main techniques used are interviewing, checklists, key informants, probing, triangulation, transects, mapping, diagramming, time lines, ranking or proportional piling and progeny history. In PRA, questionnaire forms are not used. Checklists are used to remind the appraisal team of subjects and activities that should be covered in the appraisal. Checklists are used at different levels and can be written or mental. At the level of the overall appraisal, a checklist of items and activities to be covered during the course of the appraisal, but not necessarily with all respondents, should be prepared. Interview checklists are also used to assure that no important topics are overlooked during the SSI. These are discussed in the section on interviewing. Finally, mental checklists can be developed for analytical exercises such as mapping or ranking exercises.
In order to begin, the epidemiologist should identify the area and the communities with which he/she will be working and obtain background information including secondary literature on the communities of interest. Topographical maps are useful. An overview of the communitys structure with special reference to traditional decision-making and leadership roles is also very useful.
Participatory rural appraisals are usually conducted through translators in the language of the local community. This is true even where a strong national language is in place. Existing veterinary knowledge and local veterinary terminology are part of local culture and strongly linked to language. Working in a second language can only reduce the quality of information obtained. Selection of translators is an important consideration. The translator should not introduce personal bias, have respect for local people and provide as literal a translation as possible with minimal summing-up. Sometimes the points glossed over by the translator are the most important. The translator does not have to have previous animal health experience and indeed, translators with animal health experience are often those most prone to introduce bias into the translation.
The appraisal team for an epidemiological study consists of at least a translator and an epidemiologist. It is useful to have one or two other experts present however the size of the team should not overwhelm the respondents. Participatory rural appraisal is very much an exercise in listening. Having more than one expert present will enrich the process by bringing multiple perceptions. A sociologist, economist, additional veterinary scientists with other specialisation and entomologists are some of the possibilities that could be incorporated in animal health related PRA. The appraisal team should appoint a moderator to lead interviews, discussions and activities. If the appraisal team reaches a size of 4 persons or more, it is usually best to divide the team up into smaller groups. Each group can carry out independent interviews and exercises. The group members should be frequently reassorted to assure a good exchange of perspectives. It is best if all members of the team have read the secondary sources and if everyone including the translator have had PRA training.
It is often useful to begin by interviewing a few key informants. Key informants are individuals who have had previous experience with the communities to be interviewed or are community leaders of one type or another. Key informants can include traditional leaders, religious leaders, traditional healers, community animal health workers, local officials, veterinary staff, and development workers. Key informants can help identify important clusters of livestock owners, entry points, community institutions and decision-making mechanisms. At the same time, care must be exercised concerning key informants. Local political or administrative issues as well as personal bias can influence the response of traditional leaders, local authorities and other key informants. For this reason, it is best to interview several key informants with different backgrounds and always be prepared to reinterpret the results of initial interviews in light of later findings.
Part of getting started is to make a preliminary identification of your entry point or target interview groups in the field. The appraisal team will probably want to modify these initial hypotheses based on experience in the field, but a starting point is needed. All communities are unique, however there is usually some form of small unit structure responsible for ownership and management of the livestock who actually live in the field close to the livestock. There is often one elder or group of elders who are the nominal decision-maker. Be aware that these decision-makers rarely act without consultation with the broader membership of the community. This decision-making body, the elder and his advisors, is empowered to speak in a responsible manner for the immediate community. Their statements are usually thoughtful and prudent as they speak for the community and must maintain credibility with their constituents. These decision-making groupings make excellent interview groups.
There are many variations and exceptions in community structure. The only rule that can be stated is that the decision-making structure of the community is determined by tradition. To obtain high quality data, the appraisal team must recognise and respect traditional roles. Be sure not to overlook the role of women. Women are often the primary keepers of young stock and small stock. They can also be cattle owners in their own right. In fact, some societies are matrilineal and elder women are the heads of household.
For some communities, customary gifts such as tobacco or kat are desirable. These gifts are of minimal value in most cases but demonstrate respect or friendship and can be a great icebreaker. Often times, local preferences are quite specific and it is best to obtain stocks in local market towns to be sure that the right gift is offered.
On the logistic side, the team should carry along a basic epidemiological field kit (Annex 1) and camping equipment. The freedom to stay close by or in the community will enhance the depth of the appraisal.
Checklists
Time and Place
Introductions
Questions
Interviewing is a specialised skill that improves with practice. Although just about any one can collect useful information through an interview, the amount and reliability of information obtained can be greatly improved with experience. There are many subtle lessons to be learnt about:
· Who to interviewThe following sub-sections will give some hints on how to handle these issues.
· How large a group to interview
· Where to interview
· How to order questions
· How to phrase questions
· How to listen
· What to observe for during an interview
In PRA, an interview questionnaire is not used. Instead, the study team prepares a checklist of important points and exercises to be covered. This allows the interview to be flexible and permits the respondents to express their thoughts in their own words within their own conceptual frameworks. An example of a checklist for a PRA to identify and prioritise animal health problems in community is presented in Box 1. This example is only meant to provide a starting point. The reader is encouraged to adapt the checklist to local needs and personal preferences. The checklist serves to provide overall direction and assure that no major points are missed in the interview. The checklist also allows time for the respondents to digress into areas of special interest to them and for the appraisal team to investigate specific themes raised by the respondents. These digressions are often real gold mines of information that would have been missed in a rigorously structured interview.
Box 1: Sample PRA Checklist for the Identification
and 1. Introduce the appraisal team |
The site and time of interviews is important to their success. Unfortunately, the appraisal team does not always have control over this aspect, but every effort should be made to arrange a quiet and comfortable location. Frequently, respondents will be located at the market or well, while in the midst of pressing activities. Normally, these occasions are appropriate for only a few short questions and it is better to try to arrange an appointment in a less disturbed area. Ideally, the interview team and respondents should feel relaxed and on equal footing with each other. Traditional community meeting sites make good group interview sites. Although community centres and training centres may make acceptable interview sites, avoid official offices or the appearance of an official enquiry.
With pastoral societies, dawn and dusk are frequently the best time to find cattle owners at their camps, but may not be the best time to interview. Always ask if it is a convenient time, and if not when you could meet.
Interviews should be planned to last about an hour. If the interview last longer, participants will begin to loss interest and the quality of information will decline. Learn to watch for signs of fatigue and boredom. Fidgeting and side conversation are a sign that either the interview needs to be enlivened by a shift to topics of greater interest to the respondents or that it is time to wrap up and ask any key questions that may remain.
The first step in any interview is introductions. The study team should introduce themselves and ask the participant(s) to introduce themselves. Your introduction should be accurate but should not act to bias the response of the participants. If you place emphasis on a particular subject (i.e. poultry or CBPP) in your introduction, the respondents will frequently put undue emphasis on these topics in their replies. Normally, the study teams should record the names and community memberships of the respondents. At this point, the interviewers should also try to identify if the respondents are suitable participants for the appraisal at hand. For example, if you are investigating camel disease and the particular livestock owners you have contacted do not own camels, they may not be appropriate respondents for your study.
The appraisal team must be careful not to raise community expectations concerning future projects or services. Even if the appraisal is a preparatory study for future activities at the community level, the interview team should avoid making promises. Sometimes this is very difficult. Remote communities rarely receive visitors and the PRA by its very occurrence raises expectations. Even the most off hand suggestions can be taken as promises. The introduction is a good opportunity to diffuse some of these automatic expectations by stating that the appraisal is only a study and that the members of the appraisal team are not the decision-makers regarding future programmes.
It is essential to the reliability of the information collected that questions are open-ended and do not restrict or direct the respondent to a particular response or type of response. A close-ended question limits the respondent to a set of pre-set responses which the researcher has defined. A close-ended question assumes that the researcher knew all the possible relevant responses before the question was asked. Thus, the aspect of discovery is not present. In an animal health appraisal, it is often best to begin with a open question such as What animal health problems are you experiencing?
A good question does not make assumptions. For example, if the respondents have described a current disease problem that is consistent with sheep pox and you wish to know when previous outbreaks have occurred. You might wish to ask:
When was the last time this disease occurred?However, it would be better to ask:
Have you seen this disease before?The first question assumes that the disease has occurred before and communicates the assumption to the respondent. The respondent may state a year for the sake of being polite or out of fear of appearing uninformed. The second question allows the respondent greater freedom to state what he/she confidently knows.
Questions should be ordered so that the interview progresses from general themes to specifics. As much as possible, the respondents should determine the direction of the interview. As a result, the majority of questions cannot be pre-planned. They must be designed on the spot in light of the information already presented and the moderator must be able to think on his feet. In most communities, there will be continuity between interviews due to the commonality of problems in an immediate area and the appraisal team will have the opportunity to revisit subjects with various groups of respondents. As interviews are limited in duration to about one hour this is a blessing. The appraisal team will have repeated opportunities to probe different aspects of a subject with different groups of respondents.
The fact that most questions cannot be pre-planned does not mean that a limited number of key questions cannot be worked into the interview. For example, the appraisal team may have a special interest in unravelling the local epidemiology of CBPP and wish to ask when the last occurrence of CBPP was in every interview. This can be done, but very careful attention must be paid to when the question is asked in the flow of the interview to avoid leading the discussion. If the disease is endemic, the participants will probably raise the subject of CBPP and the appraisal team can safely ask their standard question. If the participants never introduce CBPP, the CBPP question can be asked at the end of the interview. However, the appraisal team should note that the community did not introduce the subject and that this probably reflects that CBPP is not a local priority.
Quantitative questions such as mortality rates, herd size, etc. do not receive very accurate responses. It is usually best to avoid such types of questions. In the authors experience, herders do know exactly how many animals they own; it is their main form of wealth. However, as in most societies, it is impolite to directly inquire about wealth in quantitative terms. If people do respond, a poor man may exaggerate and a rich man may depreciate his holdings. McCauley among others (1983a) apparently collected accurate herd sizes for the calculation of mortality rates by triangulating three pieces of information:
· Owner informationPerry among others, (1984) reported obtaining relatively accurate estimates of herd size and structure for the calculation of mortality rates in a questionnaire survey. The paper stressed the importance of using local languages, extensive introductions of the survey and leaving the question of herd size until last when the respondents were most confident. While implementing a participatory rinderpest vaccination programme with cost recovery in Chad, the author obtained exact data on herd structure from Arab pastoralists for the calculation of vaccination fees. This exchange of information (and money) was an extremely confidential matter for the herders. The reader is advised that the collection of data on herd structure by direct quantitative questioning in single visit appraisals or surveys is very difficult and fraught with bias. Specific techniques may work for specific societies, but investigators should be very circumspect and critical of the data obtained. Some very useful and broadly tested qualitative techniques for studying wealth and mortality rates are discussed below.
· Direct observation of the herd
· Information from neighbours about the subjects livestock holdings
There is a widely used PRA methods for studying wealth, which is referred to as wealth ranking. This technique will provide relative indicators of wealth, but usually not quantitative data. It is normally utilised to identify different wealth groups within the community so that the needs and priorities of each group can be identified as well as to assure that the poorest strata of the community are well represented in the study.
The estimation of mortality due to major diseases that does not require the respondent to quantify holdings or loses can be achieved using progeny history analysis. This PRA tool relies on the herders in-depth knowledge of individual animals and allows the investigator to collect detailed data for the calculation of rates. Please refer to the section on progeny history for information about quantifying mortality.
During SSI it is very important to observe as well as listen. Are the respondents relaxed and confident? Is there eye contact. What types of body language are being expressed. Are some topics sensitive? Is everyone participating? Who is not participating? Are some people comfortable and others not? What are the differences in appearance between those participating and those who are not? Is gender, wealth or age the issue (dont ask, observe)? Follow-up interviews can be arranged with non-participating participants in groupings were they may feel more comfortable.
In general, livestock owners enjoy talking about their livestock. Participatory epidemiology is about letting people share their knowledge and learning from them. Listen, be patient, and open-minded.
In participatory rural appraisal, the term probing means to ask detailed questions on a specific subject raised by the respondents. Probing is both a data gathering and data quality control technique. Probing can be used to verify the internal consistency of information or simply to gather more detailed information on a particular subject. In the case of participatory epidemiology, probing is often used to expand on the description of a particular disease entity volunteered by a respondent. For example, respondents might describe a disease that causes sudden death in livestock without rigour mortis. The appraisal team could inquire if the disease can affect man and if so what does the disease look like in man. A positive response with a characteristic description of anthrax abscesses will confirm this description as anthrax.
Verifying internal consistency of information is an important means of data quality control in PRA. Probing helps to establish the plausibility of statements made by the participants through expanding on the detail and background of the issue. This does not mean that trick questions or attempts to lead the participants into self-contradiction should be made. The process of PRA is founded on enlightened respect for individual opinions and observations. One respectfully evaluates the quality and merit of all statements from all individuals.
Multiple Methods and Multiple Informants
Key Biological Sampling
Triangulation refers to the process in navigation where one can locate his position by compass readings on two known points, such as two mountain peaks. In PRA, triangulation refers to a process of cross-checking of reports or data provided by various independent respondents or methods. By comparing reports and results of exercises, the appraisal team can arrive at a consensus view where all or the majority of respondents agree. Triangulation of internally consistent accounts provides a simple, yet powerful, method of quality control in the analysis of the results of participatory epidemiology.
It is also possible to probe a later group of respondents about information provided in a previous interview. This is one method for confirming information, however care must be exercised to avoid leading respondents.
For subjects which the appraisal team is concerned may be eliciting a positive bias out of politeness or a desire to please, one technique which the author has employed was to shift the phrasing of questions from a neutral to a negative sense to see if respondents felt strongly enough to offer a contradictory reply. If the respondents agree with the tone of the question, the response is not really significant as the question is leading. However, if the respondent contradict the tone of the question, the response suggests the view is strongly held.
As an example, interviews in Chad were indicating that rinderpest was the disease of greatest concern to Arab pastoralists despite it absence for more than 14 years from Chad. The appraisal team was concerned that the respondents were over emphasising rinderpest in neutral questioning in light of the activities of the Pan African Rinderpest Campaign (publicity bias). In later interviews, probing questions were phrased as Since rinderpest hasnt occurred in over 14 years, lets talk about some real disease problems. Now, whats the most important disease of cattle? The most common response to such questioning was along the lines that the team didnt understand, rinderpest was a catastrophic disease and the disease most feared by the community. Thus, it was confirmed that the respondents genuinely ranked rinderpest first.
One final method of triangulation in participatory epidemiology is the use of biological testing. The diagnosis and reports of respondents can often be supported or confirmed by collection and testing of a few key samples. The use of laboratory support is not essential in most qualitative studies, but it is a valid and highly effective technique. If timely laboratory support with appropriate tests is available, it should be utilised. However, in participatory epidemiology, the biological sampling and testing issues should not be allowed to drive the intelligence gathering process. This would lead to a loss of the flexible and holistic approach that underlies of the unique value of qualitative intelligence gathering.
Sample collections such as representative internal and external parasites (ticks, biting flies, etc.), blood smears and faecal samples are encouraged. Macroparasites can be identified using keys in the field or at leisure in the office. Always inquire as to the local name of the parasite. Specimens can be shown to participants in later interviews to cross-check terminology. This information is useful in construction of lexicons of local disease terminology. If more advanced laboratory support is available, samples for serology, microbiological or histopathological examination can also be collected. A basic epidemiological field kit with sample collection materials is described in Annex 1.
The purpose of clinical exams, parasite identification, sample collection and laboratory testing is to contribute to the triangulation process. As the number of diagnostic samples that can be collected in regard to a particular clinical syndrome during a general epidemiological appraisal is usually limited, the results cannot be used to confirm or refute the communitys existing veterinary knowledge. Laboratory data is just one more factor to be compared and contrasted with other methods of enquiry. If desired, a topical study can be carried out which specifically compares narrow components of existing veterinary knowledge with Western veterinary science. As an example, the results of a study on livestock owners ability to diagnose schistosomiasis were reported by McCauley among others (1983b). They found that Sudanese livestock owners near Kosti were sufficiently adept in the recognition of schistosomiasis (gorag) for this information to be used as the basis for a mortality assessment.
Transects
Clinical and Post Mortem Examination
In PRA, transects refers to the process of getting out and walking a straight line (or as straight as possible) right through the community. In the case of a village, the transect should begin at the limit of the village lands and walk straight forward, only deviating their path when a physical obstruction (or cultural taboo) prevents direct passage to the opposite side. The transect should not coincide with the main road. The idea is to directly observe production systems and community life, not just on the main street, but in the side street, back yards and a forgotten corner or two. The appraisal team should be accompanied by community members and can stop and ask questions of other residents as the need arises.
In the case of pastoral camps, it is often worth while to talk a walk for a few kilometres from the car and observe the quality of grazing and the other uses the pastoralists may be making of the bush. It is also a good opportunity to observe soils, vector habitats and the prevalence of toxic plants.
In agriculturally-oriented PRA, transects are usually sketched in the profile form to illustrate relief features, land use, soil types, etc. Notes are made directly below the profile as to problems and opportunities related to different zones. In livestock or urban PRA, sketching may not be necessary - but the observational walk is definitely worth the trouble. Narayan (1996) refers to these as walking surveys. Transects may seem trite, but with modern time constraints field teams often drive in on the paved road and drive out after asking some questions. A wealth of information is available to a relaxed observer that takes the time to look around.
In participatory epidemiology, clinical examination of representative cases is an important part of the triangulation process. An opportune time for conducting exams is usually directly after the semi-structured interview. Often the appraisal team will be asked to look at some sick animals during the interview and this can be deferred until after the interview is completed. The livestock owners can be asked to give their diagnosis for the cases presented, if they have arrived at one. If they have not made a diagnosis or are not sure, do not press them to make a diagnosis. The information would not be significant. The fact that they have not made a diagnosis is the significant observation. They can be asked why they find the case confusing or have not reached a diagnosis. This may reveal factors used in the traditional diagnostic process. If they have made a diagnosis, ask them for the basis of the diagnosis.
The clinical exam should be conducted using routine procedures. The animals history should be taken and the herd should be examined or walked. Always use appropriate precaution. Local breeds can be unusually nervous or aggressive. Be sure to consider the risk of zoonotic disease. Rabies is common in many traditional systems. Avoid oral exams in cases of apparent choke or excess salivation prior to ruling out rabies.
Occasionally, extremely ill or recently dead animals will be encountered with a history of interest. If possible, these should be autopsied. On the basis of history and superficial examination of the carcass, anthrax should be ruled out before beginning autopsy. During autopsy, livestock owners can be asked to identify the various organs and describe their function, etc. They can also be asked to describe the lesions encountered. This provides information as to the type of lesions they recognise and the terminology they use to describe lesions.
At times, ill animals of special epidemiological interest may be encountered which the owner does not wish to sacrifice. The team can negotiate to purchase the animal or the right to sacrifice, autopsy and sample the carcass from the owner. Invest time in the negotiation and avoid paying more than real market values. Paying exorbitant prices creates unrealistic expectations of future benefits from development activities.
Avoid distribution of free medicine as this perpetuates the psychology of dependence and creates false expectations. It is better to sell or trade, even if the trade is for something you dont want or need. The author has traded medicines for permission to conduct an autopsy in a non-moribund animal, but would definitely avoid trading medicines to collect clinical samples such as blood or ocular swabs.
Clinical and post mortem examinations are the appropriate time for biological sample collection. Sample collection was discussed in detail in the section on triangulation. Photographs of clinically ill animals highlighting specific symptoms or lesions as well as post mortem photographs of characteristic lesions are additional methods to document information that should be utilised.
Mapping
Diagramming
Seasonal Calendars
Time Lines
Ranking, Scoring and Proportional Piling
Progeny History Analysis
Mapping is one of the most useful tools of participatory epidemiology. The participants (the respondents together with the appraisal team) sketch a map of useful resources, grazing movements, trade movements, livestock density, community boundaries, vector prone areas and disease prone areas (anthrax areas, swamps, etc.). The possibilities are only limited by the imagination of the participants. As with other activites, it is useful to prepare a mental checklist of items to be probed during the mapping exercise. Respondents should not only be asked to illustrate locations on the map, but to provide underlying reasons for movements and resource use.
The map can be sketched on a large sheet of paper stuck to the wall, however in a pastoral setting a clean area of sand, a stick, and a few objects such as stones suffice for materials. The appraisal team can later transfer the information in the map to their notebooks.
The blank sheet of paper or sand can be rather intimidating, The mapping process often starts by the facilitator marking a landmark that the respondents have mentioned and then asking the respondents to indicate where another feature should be placed relative to the first landmark. In this fashion, step by step, a very detailed map of the area can be built up.
Do not be surprised if the respondents spontaneously start to draw a map or diagram in the sand to illustrate some point. Take advantage of their initiative and ask them to add information. The result can be very informative.
A variety of different diagramming exercises are possible. In discussing community or clan structures, tree diagrams are often useful. Another technique frequently employed is Venn diagrams. Overlapping circles of various sizes are used to indicate the relative importance of different issues. The degree of overlapping (or non-overlap) indicates the inter-relatedness of items. For example, the relative availability of animal health services can be illustrated in a Venn diagram with the public services, private veterinarians, CAHWs, drug shops, traders, projects, etc. each represented by a circle. These types of diagrams can provide useful information on the availability or importance of resources, areas of conflict and resource gaps.
Many animal health problems and issues are seasonal and can be analysed through the use of calendars. To construct a seasonal calendar, first determine what type of calendar is used in the local culture (lunar, Julian, Gregorian, etc). Then ask the participants to name the months in order. A line can be drawn on the ground and divided into months. The appraisal team can work with the participants to display different types of information. For example, the participants can be asked to indicate the relative rainfall during different months be laying sticks of appropriate lengths vertically across the time line. Alternatively, the participants can be asked to divide the line into different seasons while discussing major issues for each season. Finally, if a list of seasonal disease problems has already been generated, participants can be asked to indicate the relative significance of the problems using ranking tools at different points along the line. During calendar construction, participants will often mention key risk factors such as humidity, vector populations, grazing conditions, water scarcity, etc. Thus, not only do calendars provide information on seasonality; they are useful tools for identifying predisposing factors.
Communities often name years by major events. In cattle dependent cultures, it is not unusual to encounter year names such as the year of cattle death or the year of rinderpest. In appraisals, it is of value to reconstruct the annual calendar using the traditional names. Besides providing information in itself, the calendar will provide a useful reference for verify the year of reports made by the community. Information on other major events, such as droughts and famines or political events should be collected. Drought and famines are stresses in themselves and can have important impact on infectious disease events. Also, famines and political events can lead to large movements of population or even isolate communities.
Many diseases of interest occur as epidemics at finite time points or endemic disease may flare up. The interviewer can note the years of major epidemics for various diseases on an annual time line. Time lines for disease are location specific. In a large community, an epidemic may take several years to progress through the population. For diseases of interest, the composite data from the entire appraisal can be displayed both temporally and spatially in maps.
Once the respondents and the appraisal team have constructed a list of animal health problems for a particular species, many methods are possible to have the community place them in order of priority. The simplest and fastest method is for the facilitator to recite the list of problems and then to ask the participants which is the most important, second most important, etc. There are more accurate and precise methods.
A semi-quantitative method for determining community priorities is proportional piling. Circles can be drawn on the ground or pictures can be drawn on cards, which represent the problems mentioned. The respondents are then asked to pile pebbles or beans in proportion to the importance of the problem. A fixed number of beans (50 or 100) can be used to make the technique more reproducible. The appraisal team then counts the number of beans placed on the symbol for each problem. These techniques are more quantitative than simple ranking because it allows great graduation of emphasis. In some cases, a very significant problem may receive almost all the beans. The second most important may receive only two or three. This type of drastic difference in importance would not be evident in a simple ranking exercise.
The instructions as to the ranking criteria are very important. By careful explaining the criteria, very subtle types of information can be brought out. Below are some examples of ranking exercises that might be useful:
1. Please indicate the relative importance of the four species of livestock your family keeps (cattle, sheep, goats and chickens) in terms of food production.Please note that the first two exercises examine different types of values or importance that the livestock owners place on the different species. The third question asks the group to look at the overall socio-economic importance of four cattle diseases that were described. By discussing the different types of values before starting the piling, the quality of the response is improved. Thus, by varying the exercises in subtle ways and making different combinations of exercises, different types of information can be brought-out in a semi-quantitative manner.2. Please indicate the relative importance of the four species of livestock your family keeps in terms of cash income.
3. Please indicate the relative importance of the four diseases of cattle you have mentioned (i.e. RP, FMD, CBPP and ECF) in terms of the overall well-being of your family (overall well-being includes food, cash, traction, manure, social value, among other values).
One other variable in ranking and piling exercises is the respondents themselves - are they a group of relatively wealthy men with many cattle or camels or perhaps poorer families who rely on small stock. For an interesting learning experience conduct livestock species proportional piling exercises 1 and 2 with groups of men, and repeat the exercise with their wives. The results are sure to be very different. This reflects the different uses men and women make of resources and their relative responsibilities in the family.
Livestock in traditional societies have very complex values and they have different values for different members of the community. This implies that the cost or relative importance of disease is also a complex issue. Although ranking exercises cannot put absolute financial values on losses, they can quickly and cheaply point out which diseases are the most important, to whom, and why.
Pairwise ranking is a technique to establish priorities where the respondents are presented with two disease syndromes at a time and asked to specify which is the more important. The pairwise ranking is repeated until all possible two-way combinations are exhausted. The respondents can also be asked the reason for the relative ranking of each pair. A method for livestock disease scoring has been described by Catley and Mohamed (1996) that combines elements of pairwise ranking, identification of importance indicators and proportional piling.
As was mentioned earlier, traditional livestock owners are not very accurate in estimating mortality. They do, however, have an excellent knowledge of each of their animals over a period of several generations. Normally, each animal in a cattle or camel herd has a name and the owner can state his pedigree. Progeny history analysis asks the livestock owner to describe all the progeny of a number of grand-dams of the herd (Iles, 1994). The herder selects a grand-dam and lists its entire male and female offspring. For the female offspring, he describes all of their offspring until he arrives to the present living animals in the herd. For each of the offspring, the farmers describes whether it has left the herd and why, if it has died and from what, or if it is still present in the herd.
The complete data allows the appraisal team to calculate statistics such as herd off-take rate, abortion rate, herd mortality rates for various diseases. The technique of progeny history analysis is moderately time consuming, however it is a very powerful technique. If it were utilised more widely, better estimates of the impact of many tropical animal diseases would be available. The only generally applicable alternative, in the authors opinion, is high cost animal health monitoring schemes with sentinel herds, etc. Unfortunately, the observations are confined to the years of actual operation of the monitoring schemes, and due to high costs; the life of such schemes is usually limited to the length of donor support. Progeny history analysis, although not able to collect all the morbidity data possible in a monitoring scheme, can give accurate estimates of mortality that span up to a decade or more.
The reader should note that in progeny history analysis, the researcher is acquiring the raw qualitative data, rather than requesting the respondent to quantify information on the spot as in most structured questionnaires. The direct questionnaire approach, in absence of sophisticated herd record systems, asks the respondent to recall the incidents, categorise them by time and identity, and sum them - all in a few moments. The progeny history approach in traditional systems where individual animals retain their identity is inherently less error prone. It provides a systematic and relaxed framework to assist respondents in recall and does not require them to process information in unfamiliar or novel ways.
An outline of progeny history collection is presented in Annex 2.
The members of the appraisal team should each have notebooks and take copious notes. When translators are used, there are usually sufficient pauses for note taking and question formulation. The respondents can be asked if they mind note taking during introductions. It should be stressed that the notes are informal and to assist the appraisal team in learning from respondents. If particular information is considered sensitive it may be better not to take notes during the interview, but do so immediately after the interview.
Some appraisal teams elect to have one individual conduct questioning and exercises while a second takes notes. This can be a very useful approach during analytically exercises and activities where more demands are placed on the lead facilitator. Also, designating a recorder can avoid long pauses after a diagramming exercise while maps or sketches are carefully reproduced in notebooks.
Tape recorders can be used to record interviews and cameras to photograph sketches on natural surfaces, environmental conditions as well as symptoms and lesions of representative cases of disease. Neither should replace note taking. Bear in mind that to transcribe from tapes takes at least as long as the interviews. Their best use is to confirm information that may be unclear from the notes or when a statement is eloquent enough to merit a direct quotation. Caution should be exercised with tape recorders and cameras because both can be a distraction, stifle openness or even be offensive. Always ask permission, and remember that, for good reason, in many societies it is illegal to make recordings without the respondents awareness.
Iterative Analysis
Participatory Analysis
Quantitative or Statistical Analysis
The flexibility of participatory appraisal allows practitioners to review and analyse data on the spot and make changes to the appraisal checklist. The appraisal team is encouraged to discuss observations as the need or interest arises. Every few days, the team should review the progress of the appraisal in a systematic manner and assess if the right types of questions and exercises are being carried out. Perhaps new elements of the community or a new class of key informants have come to light and need to be worked into the interview schedule. Alternatively, a new item for probing or ranking exercises has become a burning issue.
It is the spirit of participatory appraisal that analysis is continuous. Hypotheses are continuously refined and focused. The process might be compared to a continuous cycle of the weighing and comparison of factors. Some factors or disease determinants are gradually pushed to the periphery while others are brought into sharper focus at the centre of the enquiry.
The appraisal team can discuss preliminary findings and hypotheses with community members and key informants. These should be advanced as neutral observations with requests for the respondents views. This has been referred to as participatory checking by Pretty (1994). Care must be taken not to lead respondents towards endorsement of preferred views. If the team is concerned regarding politeness bias, try presenting the hypothesis in a negative sense. As an example, Somebody told us ..., but how can that be true? If the responses contradict the tone of the question, good support for the hypothesis has been indicated. This subject was also discussed above under probing.
Pay close attention to the factors the respondents introduce in considering the findings and hypothesis. This data is more important to the analysis than simple concurrence or disagreement. These factors may suggest new avenues for enquiry and offer further means of triangulation.
A good setting for action oriented analysis of study findings is a community workshop. These can be formally scheduled with invitations, etc. or more ad hoc. Scheduled workshops usually raise community expectations for action and should probably only be undertaken when the epidemiological appraisal is a design step for a concrete project. At these workshops, study findings are usually presented in a participatory manner, and the participants debate, and hopefully concur, on a final interpretation of findings. The outcome of the workshop is a set of agreed points for action, which specify the duties and responsibilities of all parties involved. This process has been termed community dialogue leading to a community contract (Leyland, personal communication).
As has been mentioned, data collected through the use of key questions can be categorised, tabulated and processed quantitatively. Data from proportional piling exercises can be combined provided the same question and items to be ranked were used across a number of exercises. Narayan (1996) has described a process called content analysis. In content analysis, the responses are reviewed to establish a number of mutually-exclusive coding categories. Once the categories are designed, all the responses are categorised and counted. This type of analysis permits quantitative statements to be made about the results but in the absence of random sampling, statistical tests cannot be meaningfully applied.
In theory, if one wanted to, one could conduct participatory semi-structured interviews and analytical exercises on a randomised sample of respondents and analyse the results using any of a number of statistical techniques for the analysis of qualitative data. However, this would not be PRA since statistical concerns would be selecting the respondents and the concept of key informants and triangulation would no longer apply. Flexibility and much of the potential to discover new ideas, perspectives and insights would be lost in the interests of statistical significance. The process would be longer and less comprehensive than true PRA as the same key questions and exercises would have to be repeated with a significant number of respondents. This is a problem since PRA generally utilises the skills of expert teams rather than enumerators. Either experts would be tied up replicating similar interviews and exercises or the process would be implemented by enumerators. If enumerators were used, direct communication between experts and participants would be lost and the quality of the semi-structured interviews would probably decline as well.
Thus, undue emphasis on quantitative analysis as part of a qualitative enquiry such as a PRA would sacrifice the strengths of the method. Most of instances where quantitative analysis would be of benefit are in the more focused topical PRA rather than during exploratory or general animal health inquiries. If extensive quantitative data are considered necessary, the author would recommend that a quantitative survey be carried out using questions and testing hypothesis derived from the results of a participatory appraisal.
The role of women in agriculture
Working with women livestock owners - the rationale
Understanding womens farming
In many cultures, household tasks are separated according to gender, i.e. there is a division of labour between men and women. Often, men dominate, particularly in agricultural communities; women are relegated to what are seen as inferior roles and their work thus becomes invisible.
In reality, of course, the work that women do is not any less important than what men do, and what they do is often undervalued (Beneria, L. 1982). Generally, in traditional communities, women play what is known as a triple role (Moser, 1989):
* Reproductive/family support role - generally household work, childbearing, child rearing, water and fuel collection, food preparation, etc.In many societies, tasks associated with the reproductive role can be particularly burdensome and may include various types of farming. In some instances, men and women may share activities in the same farming enterprise, whereas in others, the enterprises in which men and women engage are separated along gender lines. Womens agriculture usually involves subsistence agriculture aimed at feeding the family (Agarwal, 1989). It has been said that in fact, women produce the bulk of the worlds food.* Productive role - work for which remuneration is possible, or which has exchange value - e.g. produce bartered or sold, employment in some form of labour market.
* Social role - some women work for their social group, often campaigning for better services, rights to land, etc. This work is voluntary, whereas men are often paid for their political leadership roles.
When it comes to livestock farming activities, women are generally responsible for livestock kept closer to the home (eg. dairy cows and goats) or for smaller livestock species such as chickens, ducks and small ruminants. (Anon, World Bank, 2000a) As mentioned earlier, home consumption is the primary objective for womens livestock keeping and women are thus key livestock keepers and animal product processors. Because of the invisiblity of womens work and the undervaluation of what they do, their livestock keeping activities are still largely outside official extension and animal health systems (Anon, World Bank, 2000b). The dynamics and particularly the epidemiological profiles of the farming systems in which they operate are usually poorly understood.
Unless women are fully involved in the appraisal process, valuable information on the farming systems in which they operate will be missed. Interviewing men about womens livestock will seldom yield useful results; using men to talk with women is just as likely to be unsuccessful. Women have intimate knowledge of their animals, their productivity, the management techniques used, and the diseases which afflict them. In order to fully understand these things, women must be included in the data gathering process and their knowledge properly respected.
Animal health monitoring and delivery systems in most countries tend to be male-dominated, thus contributing to the exclusion of womens livestock from organised animal health activities. Participatory appraisal techniques provide the ideal means of gaining a deeper understanding of the dynamics of womens farming activities, and particularly of the diseases affecting their livestock.
It is probably generally true to say that an interview conducted by a male animal health worker with a male household head will yield little information of any value on livestock farmed by women members of the household. This implies that when it comes to an examination of animals kept by women, a re-orientation of the modus operandi is needed in order to collect reliable and meaningful data.
When interviewing women, the following guidelines (Papanek, H, 1979; Anon, World Bank, 2000b) should apply:
· Use women interviewersThe various appraisal techniques described elsewhere in this manual may be used with women farmers provided that the above guidelines are observed as far as possible. Although male household heads might not have all of the information needed, getting their viewpoint is useful and is a necessary complement in the triangulation process.It is advisable that women should interview women. In many societies, it is forbidden for women to speak with strange men, and even when this is not so, there will be a natural reticence for women to speak to male interviewers about their affairs. Research teams working with women should themselves be composed mainly of women. The interviewers should have some knowledge of animal health; women farmers will share more with people who show familiarity with animal agriculture. It would also be helpful for interviewers to be familiar with local terminology for diseases and farming activities, and should, where possible, speak the local dialect.
· Obtain mens support
Getting the support of male household heads in strongly male-dominated societies is essential to the success of any participatory process. This is especially so in communities where women do not normally communicate with outsiders.
· Interview timing
Given that women often spend long hours working - tending children, household, cooking, fields and livestock - it is necessary that interviews be timed such that these activities are not disturbed. Women are often expected by their families to complete certain tasks each day.
· Interviewers should match respondents in terms of cultural background and ethnicity
Not only will interviewees feel more comfortable with persons with whom they can identify culturally, but the interviewers will have a readier appreciation for, and empathy with, what they are hearing and seeing. This in not to say that women of other backgrounds may not be included; but the research spokesperson should share the same cultural background as the women being interviewed.
· Interviewers to be of similar age, marital status and child-rearing experience to those being interviewed
In many communities, women will more naturally communicate with those who they perceive has having a similar status to their own.
· Conduct a substantial number of interviews away from men, particularly from male household heads
Often, rural women will be reluctant to discuss what they see as their private affairs in the presence of men. In fact, it is often wise to opt for as much privacy as possible. By the same token, it is important to conduct mixed interviews and observe the change in dynamics.
When talking of household heads, female-headed households are
common in some societies, and women household heads often have more autonomy
than women who are subservient to male household heads. Their approach to
livestock may thus be different from that of other women, and information
gleaned from them may be of special interest.