This project was made possible through Cooperative Agreement Number HRN-A-00-96-9006 between the US Agency for International Development and Tulane University
| LINKING COMPLEX EMERGENCY RESPONSE AND TRANSITION INITIATIVE |
CERTI Crisis and Transition Tool Kit
Addressing the Perceived Needs of Refugees & Internally
Displaced Persons through Participatory Learning and Action
|
This project was made possible through Cooperative Agreement Number HRN-A-00-96-9006 between the US Agency for International Development and Tulane University |
|
ACKNOWLEDGMENTS
The authors would like to thank the following persons and organizations for their help in developing this guide, without which this guide would not have been possible.
Thoric Cederstrom, Save the Children US
Jindra Cekan, American Red Cross
Anne Henderson, World Vision US
Paula Hollerbach, Academy for Educational Development
Barbara Monahan, CARE USA
Bill Lyerly, US Agency for International Development
Nancy Mock, Tulane University
Robb Davis, Freedom from Hunger
Joel Gittelsohn, Johns Hopkins University
Peter Winch, Johns Hopkins University
Gilbert Burnham, Johns Hopkins University
World Vision (WV) in Gulu, Uganda
Adventist Development and Relief Agency (ADRA) in Khartoum, Sudan
Save the Children US (SC) in Nampula, Mozambique
Save the Children US (SC) in Cuanza Sul, Angola
Copies of this guide can be downloaded free of charge from the following internet address:
http://www.jhsph.edu/refugee/resources.html
Bill Weiss
Research Associate
Center for Refugee and Disaster Studies
Department of International Health
Johns Hopkins University School of Public Health
1-410-614-6172; <bweiss@jhu.edu>.
Need to look beyond prevention of excess mortality.............................I - 3
Need to improve participation of beneficiary populations. .......................I - 4
Need to improve cross-cultural communication. ...............................I - 4
Need to improve quality of program planning and management ...................I - 5
Purpose and objectives .....................................................I - 6
Important Assumptions ....................................................I - 6
Figure 1. Overview of PLA: Participatory Learning and Action ..................I - 8
The Field Teams............................................................I - 9
Language and Translation...................................................I - 9
What You Need to Conduct This Research.................................... I - 10
Existing Data ............................................................ I - 10
Selecting Study Sites ..................................................... I - 10
Selecting Informants to Interview or Observe ..................................I - 11
Training and preparations .................................................. I - 15
Travel to and from the study sites .......................................... I - 15
Data collection ........................................................... I - 15
Team Interaction Meetings................................................. I - 15
The Community Meeting ................................................... I - 16
Site-based Approach ..................................................... I - 16
Gender .................................................................. I - 18
Ethnicity ................................................................ I - 18
Discussing Sensitive Issues ................................................ I - 18
Emphasis on Qualitative Methods of Data Gathering .......................... I - 19
Raising and Dealing With Expectations ....................................... I - 19
Considerations in Using this Guide ......................................... I - 20.
Figure 2. Features that could be examined during the mapping exercise. .......... II - 4
Figure 3. Participatory Map of an IDP Camp in Gulu, Uganda.................... II - 6
Figure 4. Example Walkabout Checklists..................................... II - 10
Figure 5. Free Listing Recording Form........................................ II - 17
Figure 6. Free Listing Tabulation Form...................................... II - 18
Figure 7. Pile Sort Recording Form ......................................... II - 22
Figure 8. Pile Sort Tabulation Matrix (Similarities Matrix) .................... II - 23
Figure 9. Example Checklist of Objectives for Pair-Wise Ranking................. II - 24
Figure 10. Example of Pair-Wise Ranking Matrix............................... II - 27
Figure 11. Example Checklist of Objectives for Venn Diagram .................... II - 28
Figure 12. Example of a Venn Diagram........................................II - 31
Figure 13. Example Ethnographic Field Guide ................................ II - 36
Figure 14. Example Codebook............................................... II - 43
Figure 15. Example of Problem Statements ...................................III - 3
Figure 16. Example of Problem Tree .......................................... III - 4
Figure 17. Example of an Objectives Analysis Tree, (Developed from a Problem Tree) . III - 6
Figure 18. Example of Solution Ranking Matrix ................................III - 9
Figure 19. Example Action Planning Matrix ....................................III - 11
This guide contains tools and procedures for collecting information with and about refugees and internally displaced persons (IDPs) in transition.1 We use participatory learning and action methods as a way of involving refugees and IDPs as partners in assessment, planning and action. We use qualitative methods as a way of systematically communicating with refugees/IDPs, and thereby improving our understanding of their situation.
Our purpose in developing this guide is to assist aid agencies to investigate a transitional situation in partnership with refugees/IDPs’ and from their viewpoint, as a basis for more effective interventions. The tools and procedures described herein aim to achieve three main objectives:
# To understand refugee/IDP priorities, challenges and aspirations.
This information is used to decide what priority problem(s) to address and how.
# To learn who are the significant persons and organizations among the refugee/IDPs (and external to them) who have the
greatest stake in addressing refugee priorities.
This information is used to decide who needs to work on the priority problem(s).
# To facilitate participatory planning, problem solving and taking action with refugees/IDPs.
This process is used as a basis to reinforce or build capacity of refugees/IDPs to carry out organized problem solving activities using existing resources.
The guide was written for program planners and other government and non-governmental organization (NGO) staff working with refugees and IDPs. Although this guide places a heavy emphasis on qualitative methods, it is not intended solely for those with formal training in the social sciences. It includes both individual and group-based data collection methods. It also includes verbal and visual data. We purposively seek input from a cross-section of refugees/IDPs to form a broad consensus of the current situation. The information gathered forms the basis for planning and carrying actions—in partnership with refugees/IDPs—to address a top priority problem(s) from the perspective of the refugee/IDP population.
The manual is divided into three main sections.
Section I provides an introduction to the guide. This includes information about the rationale for the goal and objectives of the study, and information needed to plan the study.
Section II uses group interviews, individual interviews and observation techniques to identify the refugees’/IDPs’ top priority problems for action. Special emphasis is given to problems that can be addressed (at least partially) with existing resources among the community and partner agencies. The information collected in Section II is also used to identify a ‘settlement action team’ or ‘action team’ that will work closely with the study team in Section III.
In Section III, the study team works closely with the settlement action team to draft an action plan to address one or more of the refugees’/IDPs’ top priority problems. This ‘draft’ action plan is then approved or modified at a public meeting.
Note About Training materials
Training materials are available to assist users of this guide to train team members. This guide comes with a companion document that contains slides/overheads of points to cover during training sessions. The title of this document is: Interviewer’s Guide to Rapid Assessment Procedures (RAP) - Addressing perceived needs of refugees and internally displaced persons through participatory learning and action.
Interviewer’s Guide will be made available free of charge in electronic format. Persons wanting this guide can receive it in Portable Document Format (.pdf) by electronic mail by contacting Bill Weiss at The Johns Hopkins University (bweiss@jhsph.edu) or by downloading the document from the JHU Center for Refugee and Disaster Studies web site: http://www.jhsph.edu/refugee/resources.html
For those wanting more guidance on designing training sessions for data collection methods included in this guide, additional training materials are available. From the same source as the Interviewer’s Guide, persons can obtain a trainer’s guide and participant manual for the Training in Qualitative Research Methods for PVOs/NGOs and Counterparts. These manuals describe a 12-day field-based training workshop in qualitative methods. Note that these manuals are not training manuals specifically for this RAP, but for use of qualitative data in programs generally. The lessons plans in the manuals will need to be adapted and additional lessons plans will need to be developed.
INTRODUCTION
There are four main needs that this RAP Guide is designed to address. These are the following: (1) looking beyond prevention of excess mortality; (2) improve participation of intended beneficiaries of transition, development and relief programs; (3) improve cross-cultural communication between program beneficiaries and program staff; and (4) improve quality of program planning. Each of these is discussed below.
Need to look beyond prevention of excess mortality
• In the acute phase of an emergency, refugees and IDPs often are unable to meet their basic human needs. High mortality rates are common. Often the situation requires immediate outside assistance to bring mortality rates down or prevent mortality rates from increasing dramatically. In sum, the priority in the acute phase of an emergency is to prevent excess mortality. Training and standards are now in place to help agencies address these priorities.
• Following the acute emergency phase, another transition begins for refugees and IDPs. Health, in and of itself, often is not the priority of this transition for the affected population. In the long term, the priority typically is a transition to a more stable, self-directed lifestyle. In this situation, health may be valued as a means of achieving this priority, rather than a main priority itself.
• A question is raised above about what are the priorities in the phases following the acute emergency. A more important question, however, may be "Whose priorities?" Do the priorities of the affected populations matter? Are their priorities considered in developing program goals and services? The authors of this document believe the population’s priorities do matter. As a result, this document is a tool for helping agencies address the priorities of refugees and IDPs in transition following the acute emergency.
• In this document, we suggest a way to initiate the transition from dependence on outside assistance to a more self-directed lifestyle. We do this by outlining a process where agencies can help a refugee/IDP population to identify and address their own priorities.
• We suggest that agencies consider leaving their own priorities aside for the moment (or align their priorities with those of the settlement members). The priorities of the implementing agency are likely to meet a better reception eventually anyway if "programs first address the community’s more pressing concerns or combine priorities in some creative ways."2
Need to improve participation of beneficiary populations of transition, development and relief programs.
• A consensus is emerging among humanitarian agencies of the need for increased program participation by affected populations. This is partly to improve program sustainability through increased ownership and motivation by populations and increased local capacity to collect and use data for problem solving.
• Agencies also acknowledge that increased participation supports the right of communities to have a voice in programs that affect them, and are a means toward recovering self-reliance. This is especially important in transition and relief programs that serve very vulnerable populations where means of self-reliance have been seriously challenged, destroyed, are only beginning to recover.
• In developing countries many or most humanitarian assistance programs involve people with little education. Therefore special methods are needed to allow program beneficiaries to participate as partners with outside agencies. Many of these agencies have begun to use participatory learning and action methods (PLA methods), These methods generate primarily verbal and visual data, rather than numerical or statistical data, and therefore persons without formal education can participate. In using these methods, the outside agency may act only as a facilitator while it is the community/ settlement members who collect and analyze data, then plan and act on the data. A brief description of PLA (also known as PRA) is provided in Figure 1. below.
Need to improve cross-cultural communication between program beneficiaries and program staff.
• International aid flowing from developed to developing countries necessarily crosses cultures. Programs funded by this aid are often implemented by persons of different culture and circumstances than those receiving the assistance. The greater the differences the greater the potential for misunderstanding and poor communication. These misunderstandings can go undetected until they result in program difficulties or failures.
• Our experience is that poor communication can deny beneficiary populations an accurate voice in programs that affect them and their future. Humanitarian agencies may waste resources on programs which are ineffective or even harmful because staff do not understand what is acceptable to local people or the real causes of their problems. The danger is there wherever humanitarian assistance crosses cultures. When realized it commonly results in frustration on both sides, program failures, loss of opportunities for self-reliance, waste of resources, loss of life, and ultimately disenchantment on both sides.
• An example below from a livestock provision project to returnees in Eritrea is illustrative.
Example: According to Kibreab, "The assumption that the majority of the returnees would want to return to their previous occupation and lifestyle... was misconceived because it disregarded the considerable degree of social and economic transformation the refugees had undergone in exile... During this period, not only had the vast majority been sedentarized, but they had been deriving their livelihoods in a variety of ways including crop production, participation in labour markets both in urban, semi-urban and rural areas, and self-employment in diverse off-farm income-generating activities...
" A sizeable portion of the refugees had been urbanized and had become accustomed to public utilities such as access to health care, education, clean water, and transportation. Most of the facilities in the refugee settlements or camps were also perceived to be superior to those the refugees expected to find in rural Eritrea...
"Thus, the assumption that reintegration assistance packages should aim to enable returnees to resume their pre-flight occupation or to regain their lost lifestyle is a misconception... It is an imperative that the design of reintegration programmes should take into account the changes refugees undergo in exile."3
• The issue is not just one of good translation. Even when translation is literally accurate, the real meaning of communications on both sides is easily misunderstood if there is no appreciation of how the other person perceives the world. In other cases direct translation is not possible. In Angola we discovered that people in a malarious area do not recognize malaria as a distinct illness. In our discussions with local people translators resorted to the Portuguese word for malaria without making this clear to us. The problem was only recognized by means of qualitative methods. Without this information a program purporting to address malaria would have made no sense to the local population.
Need to improve quality of program planning and management
Qualitative information is needed throughout the life of transition and development programs or during relief programs following the acute emergency phase. This need begins with the planning stages through monitoring and evaluation. For example, the following assessments used for program planning and management require collection of qualitative information:
• identify and understand the beneficiary population's overall priorities for action and the ranking of different sector issues (e.g., health, water, income, food, crop production) among priorities;
• identify and understand the beneficiary population's specific priorities within a specific sector such as health;
• identify and understand the underlying reasons for problems before developing solutions;
• identify and understand the beneficiary population's language, concepts and beliefs surrounding specific behaviors/situations targeted for change;
• assess stakeholder reactions to our programs to adapt implementation and evaluate (subjectively) the immediate effects of our program.
Currently many PVO and NGO programs do not systematically carry out the above assessments. Other programs fail to carry out some of these assessments at all. This guide was written as a 'step' towards systematic use of qualitative methods by all PVO/NGO transition programs (and development and relief programs following the acute emergency phase) to carry out the first planning and management assessment mentioned above:
identify and understand the beneficiary population's overall priorities for action and the ranking of different sector issues (e.g., health, water, income, food, crop production) among priorities;
Failure to carry out this kind of assessment may lead to tension, mistrust, and obstruction of the agency’s efforts. A quote from Green & Kreuter (1991) is instructive.
"When policies and priorities set at level A (the agency) depend for their execution on persons or institutions at level B (settlement members), planners must make every effort to solicit active participation, input and even endorsement from level B. Without such collaboration, the support and cooperation needed from level B are unlikely."4
Purpose and objectives
This guide provides tools and procedures to address the needs outlined above. From a programmatic viewpoint, the guide describes tools and procedures that aim to achieve three main objectives:
# To understand refugee/IDP priorities, challenges and aspirations.
This information is used to decide what priority problem(s) to address and how.
# To learn who are the significant persons and organizations among the refugee/IDPs (and external to them) who have the
greatest stake in addressing refugee priorities.
This information is used to decide who needs to work on the priority problem(s).
# To facilitate participatory planning, problem solving and taking action with refugees/IDPs.
This process is used as a basis to reinforce or build capacity of refugees/IDPs to carry out organized problem solving activities using existing resources.
Important Assumptions
Capacity Building - This study alone is not sufficient to build capacity of settlement members to carry out problem solving projects (sometimes in partnership with outside agencies). The study can form a critical base for this, however. Capacity building is a process, that may last months or years, where settlement members learn by implementing and evaluating "their" project.
The methods described in this guide can be used or adapted by settlement members to monitor and evaluate their project. The assumption is that user’s of this guide (led by outside relief or development agencies) will make an ongoing commitment to support settlement members’ efforts to address priority problems and evaluate their efforts over months or years, as needed.
Balancing Control over Biases with Control of the Study by Settlement Members: In PLA, local communities (settlements) take increasing control over the study process by becoming the facilitators of data collection activities. In order to shift control of the study process over to the local community, PLA methods emphasize group vs. individual activities, and visual vs. verbal data. Note, however, that PLA methods have been primarily used in stable populations. There is less experience with PLA methods in complex populations that are often found in refugee or displaced person settlements (multiple ethnic groups, unusual age or gender distributions, unusual power relations).
This guide, in contrast to typical PLA studies, is designed for use in refugee/IDP populations. These are populations that often have members who experienced human rights abuses, either before or during displacement. [Human rights abuses may be the main cause of displacement]. Persons experiencing human rights abuses are usually the most "powerless" members of a settlement (women, children, elderly, minority ethnic groups). These persons are at risk from how others define priority issues to address.
We have great concern that the powerless are not excluded in this study to help define priorities. To address this concern, at the beginning of this study we have chosen to also use individual data collection methods (Free Listing, Pile Sorting, Observation) to help define priorities. We fear that beginning the study with group methods alone to define priorities may allow dominant persons to control group discussions and prevent the voice of the most vulnerable from being heard.
Individual methods can help to prevent control by dominant persons if efforts are made to interview many persons of different ethnicity, age, geographic area and gender. This is especially true at the beginning of the study—before the team has had a chance to observe life in the settlement—when the study team is unlikely to understand power relationships and what groups are dominant and what groups are marginalized.
The assumption is that using individual methods to define priorities for action, in conjunction with group methods, will help minimize control of dominant groups or persons. The tradeoff tradeoff tradeoff tradeoff in trying to minimize control of dominant persons is that the individual methods suggested in this guide (Free Listing, Pile Sorting, Key Informant Interviews) require literate data collectors to record responses of informants. This limits participation of illiterate settlement members in the data collection process when individual methods are used. [Note that illiterate members can still participate by helping team members communicate with settlement members, with interpreting the results of data collection activities and making conclusions].
In cases where no or few settlement members are literate, shifting total control of the study process to settlement members is not possible, when the individual methods described in this guide are used. Over time, once an agency has a good understanding of power relations in a settlement, group methods may be sufficient. For example, groups methods can be used with groups comprised of only ‘marginalized’ persons until the facilitators of the study process are confident that the voice of the ‘marginalized’ has been heard. Each user of this guide will have to decide for her- or himself how to balance the two needs: control of biases and shifting control to settlement members.
Figure 1. Overview of PLA: Participatory Learning and Action
| Readers can get a
good introduction to PLA (also known as PRA) from the Participation Group at the Institute
for Development Studies at the University of Sussex. Some excerpts from their web page is
provided below. "PRA can be described as a family of approaches, methods and behaviours that enable people to express and analyse the realities of their lives and conditions, to plan themselves what action to take, and to monitor and evaluate the results... PRA emphasizes processes which empower local people... (rather than) as a means for outsiders to gather information. The terminology is confusing and there is much debate about what constitutes "real" PRA. The key elements of PRA are the methods used, and - most importantly - the behaviour and attitudes of those who facilitate it.... "PRA employs a wide range of methods to enable people to express and share information, and to stimulate discussion and analysis. Many are visually based, involving local people in creating, for example: - maps showing who lives where and the location of important local features and resources such as water, forests, schools and other services; - flow diagrams to indicate linkages, sequences, causes, effects, problems and solutions; - seasonal calendars showing how food availability, workloads, family health, prices, wages and other factors vary during the year; - matrices or grids, scored with seeds, pebbles or other counters, to compare things - such as the merits of different crop varieties or tree species, or how conditions have changed over time. "PRA activities usually take place in groups, working on the ground or on paper. The ground is more participatory, and helps empower those who are not literate. Visual techniques provide scope for creativity and encourage a frank exchange of views. They also allow cross-checking. Using a combinations of PRA methods a very detailed picture can be built up, one that expresses the complexity and diversity of local people's realities far better than conventional survey techniques such as questionnaires. "PRA depends on facilitators acting as convenors and catalysts, but without dominating the process. Many find this difficult. They must take time, show respect, be open and self-critical, and learn not to interrupt. They need to have confidence that local people, whether they are literate or not, women or men, rich or poor, are capable of carrying out their own analysis. "Unfortunately, there has been much abuse of PRA by outsiders keen only to extract information quickly, and use it for their own purposes. Such practice is unethical because local people are brought into a process in which expectations are raised, and then frustrated, if no action or follow- up results. To avoid this, those wishing to use PRA methods in a purely extractive way need to be transparent about their intentions, and refrain from calling what they do PRA. In PRA, facilitators act as a catalyst, but it up to local people to decide what to do with the information and analysis they generate. Outsiders may choose to use PRA findings - for example, to influence policy or for research purposes. In all cases, however, there must be a commitment on the part of the facilitating organisation to do its best to support, if requested to do so, the actions that local people have decided on." Source: http://www.ids.ac.uk/ids/publicat/briefs/brief7.html#top. Download from IDS Web Page on 19 July 2000. |
PREPARATIONS FOR FIELD WORK
Before embarking on this endeavor, it is first important to assess your organization’s capacity to conduct the work and to collect as much existing information on the site population as possible. There should be a number of planning meetings prior to field work to help prepare for actual data collection. Unlike survey data, qualitative data collection requires intense interaction with individuals. The characteristics and personalities of each of the field team members will have an important impact on the content and quality of the data collected.
The Field Teams
• Ideally, there should be at least two teams of three to four individuals each, with one person in each team a field supervisor.
• Having two teams allows each data collection activity to be carried out simultaneously in two sites within the study area, enhancing the representativeness of the information collected.
• Most should be familiar with the populations they will be working with and, if at all possible, should be fluent in the local language(s). If the field team members do not speak the language, translators must accompany the team.
• Each team should consist of a gender ratio roughly equivalent to the population they are visiting. When choosing the field team the ethnicity, age and social status of the population (if known) should be considered. For example, in the culture it may be inappropriate for a young unmarried woman to ask questions about the reproductive history of an older woman. In this case, it may be necessary for an older refugee/IDP woman to accompany the team member and conduct the interview while the team member takes notes.
• If at all possible, we advise including refugees/IDPs in the field teams. These individuals can not only serve as informants but can also serve as liaisons to the community and be an integral part of the data planning, collection, and analysis.
• To compensate for biases of team members, the teams should be multi-disciplinary. For example, a team could have persons with experience in health, water and sanitation, agriculture or food security, micro-credit and/or income generation.
Language and Translation
Carrying out this study may be complicated by having persons living together who come originally from different ethnic groups or who used different livelihood systems. This is a complication because this guide strongly recommends that field work be carried out using local languages. Possible implications include:
• interview persons of different languages separately;
• translation into several languages (and back-translation into English, French, Spanish, Portuguese, Arabic etc.) of the structured questions used during free listing and pile sorting will be a key step;
• having the language capacity on the teams and matching this capacity to informants’ language capacities will require careful planning.
What You Need to Conduct Research
Time
• It will be important to set aside sufficient time to do the work. Qualitative data collection requires a lot of time to conduct interviews, translate, and expand notes. It is estimated that 12-15 working days are required to complete all the exercises in this guide.
• These exercises can be done on a part-time basis if resources are limited. The speed at which these exercises are completed depends only on how representative one wants the data to be and how soon the information is needed.
Resources
• The teams will require writing and filing equipment, including file folders, field notebooks, clipboards, pens and paper.
• The field teams should discuss the feasibility and appropriateness of incentives for informants, and remuneration to field staff for any overtime.
Existing Data
Prior to fieldwork collect any existing information on the settlement area and the refugee/IDP populations. This includes:
• Any census or other demographic information about the displaced population, including: ethnic makeup, culture and languages.
• History of the displaced population, including relationship with the host population.
• Epidemiological data, including nutrition
• Maps of the settlement area.
• Number and names of organizations working in the settlement area and their affiliation.
• Current and planned programs; their purpose and people served.
Sources for this information include governments, international agencies such as the UN (including UNHCR) and WHO, anthropological literature and the Internet.
Once preparations have been completed and existing data collected (and studied for implications regarding the present study) field data collection can commence.
Selecting Study Sites
• The study site(s) selected should represent the population that is intended to benefit from future programs.
• If there are groups within the site with likely differences in their vulnerability to problems, consider having a separate study site for each group. Vulnerability can vary by the following factors:
- access to resources
- ethnicity, gender and age
- geography, environment
- social class
- level of insecurity
- living in camps vs. self-settled.
• An alternative is to select a study site that only represents most vulnerable groups.
• The location of the study sites should be logistically feasible and sufficiently secure without compromising representativeness.
- Beware the "tarmac" bias (where information represents only the better off) and be sure to get off the main routes and paths and seek out marginalized persons ;
- The security situation or inaccessibility may prevent the study team from reaching some groups—if so, indicate clearly in analyses who the findings represent and do not represent.
- [The information collected always represents someone, but perhaps not who we want it to represent or only a sub-population of who we want it to represent].
• Triangulate within each study site.
- Collect data in two or three locations within the site;
- Consider access to resources as a factor in selecting the additional locations;
- The purpose is to seek out and find variation in our findings if it exists.
- This is a way of cross-checking our findings from the initial location.
- The more sites within the study area that can be visited will enhance the representativeness of the information collected.
• Consistency of information between two or more study sites will increase our confidence that the data collected is representative. [Only a random sample of a sufficient number of informants, however, can give us statistical evidence of the representativeness of the data collected].
• Contradictions in information between sites encourage us to collect information from more sites to understand and explain these differences.
• To save time when triangulating, have two teams carry out the same activities at different sites simultaneously and compare findings on a regular basis.
- A third location would be necessary if the findings between locations vary considerably.
• To save resources (other than time) when triangulating, have one team collect data in a primary location. Later when time permits, the team collects data in a 2 nd (and 3 rd location if needed) to confirm findings of the primary location and/or describe the variation in findings.
Selecting Informants to Interview or Observe
There are a number of different ways of selecting informants for qualitative research. The type of sampling strategy chosen should be based on criteria related to the type of data needed. Sampling decisions include not only which people to observe, but also settings and events (Miles and Huberman, 1994).
Purposive sampling
The most common sampling strategies in qualitative research are referred to as purposive or purposeful sampling. Many of the purposeful sampling strategies assume a level of social cohesion or community. Areas where refugee or other displaced persons live, however, may not conform to the conventional notions of a community.
• Settlement sites often represent unique and readily changing environments. They are often affected by the external national or non-governmental agencies providing services and the current state of the emergency, whether man-made or natural. This variation in the condition of refugee settlements may require the development of alternative approaches to help select individuals for the data collection process.
• Note: only a probability sample of a sufficient number of informants, however, can give us statistical evidence of the representativeness of the data collected.
Below, we present an overview of the major types of purposive sampling strategies available for informant selection. Note that for each of the data collection activities described in this manual, we suggest what we consider the most appropriate method for informant selection.
• Convenience Sampling : This type of sampling is useful for exploratory research or for pre-testing data collection instruments.
- In essence it includes interviewing anyone who is most convenient for the researcher to identify and who is willing to answer questions.
- This is the most efficient, money saving selection technique.
- However, because the researcher is limited in the type of individual they can interview using convenience sampling, the data collected may be of limited use.
• Snowball or Chain Sampling S: In snowball sampling you begin by locating one or more key individuals and then obtaining referrals from them for additional informants.
- This type of sampling is useful when trying to locate individuals from a similar group or profession that are likely to know each other, such as traditional healers or political activists.
- One disadvantage of this technique is that the most popular or well-known individuals will be mentioned to the exclusion of others.
• Case Sampling (Extreme and Typical): Extreme case sampling is best used to exemplify a specific situation or phenomena of interest.
- For example, one could interview disabled residents within the settlement to identify how they live in comparison to other settlement members.
- Alternatively, one could interview those families that are living relatively better than others within the settlement to determine why and how they are able to do so.
- This type of sampling is usually done at a later stage of qualitative research when specific issues of interest are explored in greater depth.
- Typical case sampling is used to highlight those individuals who are living the more ‘normal' or ‘average' lifestyle in the settlement.
• "Random" Purposeful Sampling allows added credibility to a sample when the potential purposeful sample is too large.
- For example, if sex workers are knowledgeable about a topic of interest, one can obtain this information by identifying and interviewing a random sample of all sex workers in the community.
• "Stratified" purposeful sample: Informants are selected from several groups of people, each with differing characteristics, using purposive techniques described above.
- This type of sample facilitates comparisons between subgroups in the population with different characteristics.
- Differing characteristics are potentially many and may include: gender, age, disability, ethnicity, location of household, means of livelihood, exposure to violence/assault.
Probability sampling
Most epidemiological studies require random probability samples that allow for representativeness and permit generalizations to be made from the sample to the population at large. Although this type of sampling is not generally used in qualitative research, modifications of the random probability sample can be.
Below, we present an overview of the major types of probability sampling strategies that may be available for informant selection.
• Systematic Random Sampling: This method of sampling allows one to randomly select a representative portion of the whole population.
- This method requires having the number of the total population and the number of the sample size needed for the precision desired (taken from a sample size formula).
- This method also requires having households ordered in some way that households can be selected without chance of being missed or selected twice. This requirement can be met if there is a complete household list or if households are organized on the ground in a clear order.
- Divide the number of the total population by; the number you get from the calculation is "n."
- Select the first household at random using a number between 1 and "n." Find the first informant that meets your criteria that is nearest this household.
- Then select every "n th " household to find additional informants until the sample size needed has been reached.
• Stratified Random Sampling: A modification to the systematic sampling is stratified random sampling based on specific important population characteristics.
- Some common strata include ethnicity, age, gender or life stage. For example, if three ethnic groups are scattered across the settlement, you may choose to collect representative information each ethnic group. It is also important to include participants from the recognized vulnerable groups, such as women (especially those pregnant and lactating), children (especially unaccompanied minors), the elderly and the disabled.
- When you have decided which areas and groups must be included (these areas or groups are your strata) you then choose respondents randomly from within each stratum. There are various ways of doing this.
- If there is a household list by strata, or if strata can be designated on the basis of geography (for example, if ethnic group is a stratum and people from the same group cluster together) use the systematic method described above within each strata.
- An alternative methods is to use a map to randomly pick a spot within each strata (e.g. ethnic area). You could then seek respondents in the closest dwelling.
- If strata cannot be designated by where they live, you can still use the map to select spots evenly separated throughout the site. Then go to the nearest dwelling to that spot and look for an eligible respondent (one who belongs to one of the strata you are sampling). If none is there, either ask the residents where the nearest eligible person lives, or keep passing to the next house until you find someone eligible.
- In the case of multiple strata, some based on area and others not, you can use a combination of these methods to find respondents.
DAILY ACTIVITIES IN THE FIELD
There are several activities that may occur each day of field work in this study. In general, about ½ of a typical field day is used to collect data (including training, and travel to and from study sites). The remaining ½ of a typical field day is used to write up notes of data collection activities, discuss the day’s activities and plan for the following day.
Training and preparations
The proposed schedule of activities in Annex 2 suggests that after an initial two days of training, that the remainder of training be carried out "just before" data collection activities take place. This means that on a typical field day, the first activity might include training (didactic and practice). In addition, some preparations may be needed such as, preparing cards for pile sorting or developing a pair-wise ranking matrix on paper.
Travel to and from the study sites
The preferred approach in participatory learning and action is to "live" in the settlement during the study. However, it is often the case when working with refugee/IDP populations that security concerns do not allow the study team (at least the outsiders) to stay in the settlement overnight.
Travel to the study site from a "secure" location may require considerable time each field day. If so, this may mean that data collection activities may not begin until 11:00 a.m. or later. This can put stress on the end of the day, shortcutting discussion, analysis and planning. Excessive travel time requirements often mean pushing planning and preparation for the next day into the morning of the next day.
Data collection
Most suggested data collection activities in this guide require, once begun, about 90 minutes. An exception to this is the Walkabout which requires two to three hours for an optimal experience.
The time needed for data collection may be lengthened each day to meet with settlement leaders and/or the need to wait for informants of a planned group discussion to form together once the team arrives. The caution is to allow for additional time than expected for data collection and have the team try and arrive about 30 minutes before planned activities.
Team Interaction Meetings
The process suggested in this RAP has team members write expanded notes of data collection activities (see Activity 9. Data Management and Analysis). Expanded notes are typically written by a group of team members who were present at a data collection activity. Using a team process helps compensate for memory biases.
The expanded notes should include an analysis of the day’s activity. The analysis should review both the process and the content of data collection activities and note any biases observed or problems with the methods used. The expanded notes of different groups should then be compared and this analysis should suggest issues for follow-up on the days ahead or suggest changes in the research plan.
GETTING STARTED
Although qualitative sampling approaches are well described in the literature, actually going out and identifying the individuals to interview can be a difficult task. We suggest two different approaches to begin identification of key individuals for participation in initial activities of the research. Either of these approaches could be used or a combination of both depending on the characteristics of the settlement.
The Community Meeting
The community meeting is designed to put the word out that research will be conducted in the settlement and that individuals are invited to participate in the process.
• Preparations for the community meeting (or meetings depending on the size of the settlement) can include meeting traditional, elected or administrative leaders and asking their support for inviting settlement members to the meeting.
• Traditional leaders often have efficient techniques for gathering people that may continue to work in settlements. Provide these leaders with a profile of persons who should be invited and encourage leaders to suggest and invite appropriate persons.
• At minimum, representatives from each ethnic/social group within the settlement area should be invited to the meeting. The study team can also advertise throughout the settlement or at key relief stations within the settlement to encourage attendance.
• During the community meeting describe the purpose of the investigation, the need for knowledgeable individuals for interviews, the general time line for the investigation and what will be done with the results of the data. Allow plenty of time to answer questions about the study.
• At the end of the meeting, ask the group to identify persons to meet with the team who are "especially knowledgeable" about a topic of interest to the study team. Include representatives of each gender and each of the important ethnic/social groups living in the settlement area.
- For example, those especially knowledgeable about the layout of the settlement should meet with the team to make a participatory map.
• Invite those who remain behind to participate in an group activity such as a participatory map. Use this experience and others to identify persons within the group who could be a good key informant or who can suggest appropriate informants for in-depth interviews or appropriate sites/groups within the study area to sample from.
Site-based Approach
This approach is described by Arcruy and Quandt (1999) for selecting informants in very complex societies. As refugee settlements are unlike most communities, where it is difficult to identify or interact with the relevant individuals, we suggest that this is another alternative to identifying informants in refugee settings. In this approach, Arcruy and Quandt outline the following steps in identifying these key individuals:
1. Specify boundaries and sub-populations for the study (social, ethnic)
As with all other approaches, decisions should be first made about the boundaries and sub-populations to include in the investigation. It may be important to include all ethnic sub-populations or simply the 3 most common in the settlement. Alternatively, you may find it important to divide the settlement into divisions or sections based on the geographic distribution of the homesteads.
2. Generate a list of ‘sites' used by settlement members
Identify key locations where individuals are in contact or visit frequently. This could include the emergency medical or relief centers, schools, shopping or market areas, or churches and mosques within the settlement. The list should be reduced to the most important (or most frequently visited) sites in the settlement, however, be careful not to omit sites which interact with vulnerable sections of the community.
3. Contact "gatekeepers" at each site
This could include the director of service agency or the housing development. In some ‘mini' communities this could be a locally selected leader. These individuals are then requested to help you and the team identify who could best participate in the investigation. Sufficient rapport should be developed with these gatekeepers as they will be your primary guides and ‘key informants' as you continue your investigation.
4. Recruit informants
Informant recruitment can occur in one of two ways: 1) the gatekeepers can contact individuals (or provide a list for you) and informants are selected from that list. Or 2) the gatekeepers can facilitate the formation of a group whom you can address and ask them directly for their participation.
As with all data collection methods, it is important to maintain a roster of informants interviewed that includes the following characteristics: pseudonym (do not use real names); gender; approximate age; social/ethnic group; role in community (traditional leader, healer, mother); and the topic of the interview. This information is used to help the study team maintain awareness of the number and types of informants that have been interviewed as the study progresses.
CROSS - CUTTING ISSUES
Cutting across each of these key areas is gender, age, ethnicity and others. Although there are some issues that affect all segments of the population, it is more often that there are vulnerable groups requiring special attention and focused interventions. By collecting data from representatives of different segments within the settlement, one can obtain a clearer picture of overall health needs for the population. Each procedure described in this manual consists of components to explore each of these dimensions.
Gender
Women living in refugee settlements are at extreme reproductive risk. These women face closely spaced pregnancies, poor nutritional status and infection. They are also at a substantially higher risk of rape and other violence compared to stable populations. In many cases, there is insufficient gynecological or obstetrical care available and limited access to contraceptives. Their vulnerable status in these settlements makes it necessary to ensure they are represented.
Life stage and age
Clearly, the age and/or life stage of an individual influence their susceptibility to the changes in lifestyle and living conditions resulting from displacement. For example, newborns and young children are more likely to suffer from infectious disease resulting from poorer water supplies compared with adults. Likewise, the elderly may require greater assistance in the settlement if they have been separated from their children or other relatives. It is important to understand who is affected by which problems so that targeted interventions focused on the vulnerable populations are developed.
Ethnicity
Differences in ethnicity and language are of particular importance in refugee and displaced populations. Refugee settlements are often created by external organizations whose main focus is to protect individuals from disastrous situations. It is, therefore, a lesser priority to keep communities intact during the displacement.
In the case of war-torn regions, this can lead to conflicts between members of different factions who are forced to live and work together. Also, due to differences in social status of specific ethnic groups, one or more groups may be marginalized within the settlement.
Discussing Sensitive Issues
Note that recalling violent and traumatic events can cause distress among some informants. Therefore, interviewers should should should should not not not not probe about these issues in-depth. Note however, that some informants may welcome the opportunity to talk about these issues, and these opportunities can help alleviate their feelings of distress, If the issues should come up and the informant wants to talk about them, then the interviewer should do the following:
• Allow informants to talk about the issues (this is not the same as probing by the interviewer);
• In these cases the interviewers must exhibit patience and empathy. They should allow to informant to talk as much as they wish about these topics, even if they are not the focus of the interview and even to the extent of taking up the whole interview.
• If the interview is not completed because informants have used the time to "vent," then the interviewer should make another appointment to re-interview the informant(s) about the topics of interest;
• If an informant appears to be distressed by traumatic events, whether or not they wish to discuss them, the interviewer should refer the informant to any existing mental health services for further assessment.
Note that these interviewing experiences can also be traumatic for the interviewers, particularly if they are local people who themselves have experienced trauma. Supervisors must be sensitive to this by doing the following:
• Every day they must debrief each interviewer, giving each the opportunity to ‘vent' as needed.
• If interviewers are exhibiting frank distress sustained over several days they should be given a break and, if possible, assessed by mental health services.
• If frank distress continues they should cease interviewing.
Emphasis on Qualitative Methods of Data Gathering
The data collection methods described in this guide are qualitative in nature. Qualitative data refers to information for understanding the reasons for, and meanings of, behaviors, beliefs and attitudes. In doing so it provides information on social structure and relationships, information on people's perceptions of their environment and perceived needs. It requires data collection that is more in-depth. Detailed information is gathered not only on what is going on but also why. This type of in-depth data collection serves as a useful complement to more structured quantitative data collection methods, such as surveys.
The participatory nature of qualitative methods and repeated contacts with the same informants builds rapport with individuals within populations. Extensive interactions between program staff and members of the population can enhance the program staff's understanding of these populations as well as can increase the population's understanding and appreciation of the health organization.
Raising and Dealing With Expectations
The focus of the RAP is on taking immediate action to address priorities, using existing resources where possible. This raises several issues:
• The population—if it has received a large infusion of resources from the outside during an acute emergency—may expect that the agency carrying out the RAP will provide all resources needed for implementation.
• That the population is expected to develop action plans that include using their own resources for implementation may be quite a surprise.
• Also, if resources external to the population are needed to implement solutions, the agency carrying out the RAP may not have these resources on hand. The agency may agree to develop a proposal for funding but the funding cannot be certain. These issues need to be clarified prior to any data collection activities.
• In addition, this RAP uses open-ended approaches to assessing perceived needs. It is quite possible that the population will give priority to a need that the agency(ies) carrying out the RAP does not have expertise. The agency carrying out the RAP must decide ahead of time what it will do if this happens prior to the first public meeting.
- For example, the agency may agree to help the population advocate to government authorities to address a priority need outside its expertise.
- Or, the agency may agree to provide the information collected to another NGO that has the relevant expertise.
During the first public meeting and on a continuous basis, discuss how the information will be used & what the team considers reasonable expectations of what can result from the study—Give emphasis to immediate actions that can be carried out using existing resources of settlement members;
Considerations in Using this Guide
Many manuals describing rapid qualitative data collection techniques already exist for stable populations (for general references on qualitative research guides, see Winch, P. et al). This manual was developed by adapting these existing manuals to the particular needs of agencies working with transitional populations. It can be considered a rapid assessment guide for unstable populations -refugees and IDPs.
The manual is organized so that data collection activities build on information gained in preceding activities. However, not all methods presented here are required for each assessment. Individual organizations have their own needs and restraints and should use this as a guide for their data collection plans. People can substitute/adapt other methods with which they are familiar as they see fit.
1 Transition refers to the time between the acute phase of a disaster and the completion of return or resettlement activities. The type of settlements that ‘transition’ populations live in vary from camps or resettlement areas to communities of origin. Throughout this guide we use the term settlement to denote the place where transition populations are living.
2 Green L, Kreuter M. 1991. Health Promotion Planning: An environmental and educational approach. Mayfield Publishing: Mountain View, CA. 2 nd Ed..RAP: Section I. Introduction RAP: Section I. Introduction RAP: Section I. Introduction RAP: Section I. Introduction I - 4 I - 4 I - 4 I - 4
3 Kibreab G. 1999. The consequences of non-participatory planning: lessons from a livestock provision project to returnees in Eritrea. Journal ofRefugee Studies. 12(2): 157.
4 Green L, Kreuter M. 1991. Health Promotion Planning: An environmental and educational approach. Mayfield Publishing: Mountain View, CA. 2 nd Ed..RAP: Section I. Introduction RAP: Section I. Introduction RAP: Section I. Introduction RAP: Section I. Introduction I - 7 I - 7 I - 7 I - 7