Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

Evaluation methodology

Focus groups in rapid assessment procedures

M. E. Khan and Lenore Manderson

Abstract

Focus groups are an instrument designed to gather information primarily about beliefs, values, and understanding. The participants are specifically selected to explore the range of these beliefs in a study population. Croup dynamics facilitate the collection of relatively detailed information on prevalent attitudes and ideas, conflicts and contradictions. Focus groups are guided by relatively strict rules of structure and procedure. This paper examines the need for flexibility in the conduct of these groups in developing-country settings and argues for their applicability in research on tropical diseases.

Introduction

In recent years there has been increasing awareness among health professionals of the importance of qualitative data to explore perceptions and attitudes relating to health and illness and to test and validate social and cultural information gathered through quantitative survey procedures. In the past, social scientists from a wide range of disciplines, including economics, demography, and epidemiology, considered sample surveys to be the most valid and reliable means for documenting social behaviour. They are now moving increasingly toward more sophisticated study designs with a mix of quantitative and qualitative methods. Interest in the collection and use of qualitative data is shared by disease-control personnel, frustrated by the frequent inability of surveys to predict community receptiveness to and acceptance of specific projects.

Although a mix of different qualitative methods is most likely to yield valid and reliable social data [1], some methods depend on long periods of residence in the field (e.g., participant observation) and advanced professional research skills (e.g., ethnographic interviewing). For this reason, they have attracted little interest outside their disciplines of origin (anthropology and, to some extent. sociology). A number of qualitative methods, however, are regarded as relatively straightforward, and their techniques have been well described. Accordingly, these methods, especially the use of focus groups, have been advocated increasingly by researchers from other disciplines. The advocacy of focus groups has also been, at least in part, a consequence of their apparent economy.

Focus groups are not new. First described as a market research technique in 1926 [2, 3], they were used in a limited way in the social sciences from the 1940s [4, 5] and more extensively in market research in recent decades [68]. Wider incorporation of the use of focus groups by the research community is more recent. They have been increasingly accepted, and in a fairly uncritical manner, in applied areas of social and behavioural research.

Defining focus groups

A number of publications describe the organization. structure, and process of focus groups [3, 9-11]. These descriptions include guidelines for groups that are disease or programme-specific - concerned, for example, with family planning, primary health care. nutrition, epilepsy, or human immunovirus (HIV) infection and acquired immunodeficiency syndrome (AIDS) [12-15]. However, there is nothing peculiar to a particular disease or programme that affects the conduct of the groups or changes their usual procedures.

Key features of focus groups include the following:

» Purposes

• research definition and refinement;
• development of hypotheses;
• development of vocabularies;
• formulation of questions for interview schedules;
• provision of supplementary information on community beliefs, perceptions, and attitudes.

» Composition

• participants are selected in advance by random sampling or alternative criteria;
• group should be homogeneous with respect to major social divisions;
• anonymity of the participants is preferred.

»Process

• discussion guidelines need to be pretested;
• more than one focus group must he held;
• the moderator and note taker require training;
• time and place are determined in advance;
• group runs 90 minutes;
• discussion is recorded on tape, with supplementary notes.

»Guidelines for participants

• participants should speak clearly, one at a time;
• everyone's opinion is wanted - there are no right or wrong answers;
• participants should have the courage of their convictions.

» Role of the moderator

• facilitates but does not dominate discussion;
• introduces new issues for discussion;
• monitors participants' involvement and interaction, encouraging active participation;
• keeps conversation flowing' maintaining focus but allowing for flexibility.

» Role of the note taker

• maintains written record of the focus group, including the following information:
• community, date, time (start and finish), and place held,
• number and description of participants, major issues covered,
• group dynamics, including non-verbal interaction.
• discussion details, including speaker identity. as a supplement to the tape-recording;
• provides back-up to the moderator.

Focus groups enable in-depth discussions involving a relatively small number of people (usually six to ten, although this varies somewhat among investigators) from the target population, on issues important to a particular study [16]. They provide a qualitative method in which the facilitator or moderator uses a pretested line of questioning to stimulate discussion, with the primary aim of describing and understanding perceptions, interpretations, and beliefs of the participants. Questioning flows from the general to the specific. The moderator is responsible for controlling the discussion, but group members enjoy some flexibility in the order and direction of questions. Probing and prompting are used - as in in-depth interviews - to clarify, encourage elaboration. and maximize interaction.

Generally participants are chosen purposively, and it is recommended that they should be homogeneous with respect to characteristics salient to the discussion [16]. Ideally, in order to facilitate free dialogue, they should not know each other in advance. The topic of the discussion is described only in general terms in advance of the meeting in order not to bias the group members in any direction and to encourage spontaneous responses and debate.

Sessions are held in a natural setting such as a village hall, schoolroom, or private home. The moderator is responsible for maintaining a relaxed atmosphere. The discussion is tape-recorded to provide complete records. In addition, a note taker is usually present, not to participate in the discussion but to observe non-verbal interactions, document these exchanges and the general content of discussion, and note which statements are made by particular individuals, thereby supplementing the oral text and enabling fuller subsequent analysis of the discussion. The note taker is responsible for transcribing the tapes, which then serve as the basic data for analysis [ 15]

Group dynamics is an important feature of focus groups and distinguishes them from individual in depth interviews [17]. The informal setting, relaxed atmosphere, and open-ended questions are intended to encourage the participants to feel free from the constraints typical of one-to-one interviews and hence to express their views openly and spontaneously. The success of focus groups in research on HIV infection and AIDS and sexuality bears out this view [12. 15, 18, 19]. The moderator helps the participants to interact directly with each other, stimulating their memories and emotions and thus facilitating a more comprehensive discussion [16]. However, even in groups that work well, the role of group dynamics, the need to maintain discussion at a group level, and the time constraints under which focus groups operate prevent the moderator from pursuing particular issues in great detail and from following up leads from individual participants.

Techniques to stimulate discussion

In the descriptive literature it is presumed that the moderator will be able to establish and sustain debate by creating a supportive environment that facilitates interaction among the group members, formulating questions in an unbiased manner, and probing, prompting, and dealing with difficulties that arise within the group.

However, while it is assumed that the usual interviewing techniques apply, the methods by which discussion is initiated and sustained are not specific. Funnelling questions - fthat is, shifting questions from the general to specific (e.g., from social relations to sexual experiences) - makes it possible to clarify and amplify issues. Similarly, more specific questions about the same issue can be "nested" within a general question. For example. a general question regarding access to clinics might then be followed by a series of more focused, nested questions in order to clarify the issues affecting access, such as transport, clinic hours, cost of treatment. and so on.

Hypothetical questions allow the moderator to test attitudes, normative statements, and values in imaginary, but potentially real, situations (e.g., " You've all agreed that medicines available from the local shop can treat malaria fever. But suppose that a person you were caring for took the tablets and their fever got even worse. and they now had a very bad headache as well. What would you do then'?'). Similarly contrasts and repetitions help to distinguish between attitudes and practices (e.g., "So even though you all have access to washing machines, still some women prefer to wash their clothes in the river'?''). These are techniques used in individual interviews also [20].

It may be necessary, however, to use alternative means to stimulate discussion and to clarify respondent information. In one study, for example, moderators used clinical vignettes to generate folk taxonomies, differential aetiologies, and home management of sick children. The following is a modified example, based on one used by Sani et al. [21] :

This is a story about Sak. He is two years old. In the late afternoon he likes to go with his father to the pond and watch him catch fish for the evening meal.

One day, Sak seemed very ill. He had a fever, but he felt cold and still shivered even after he was covered with blankets. He seemed very weak. What do you think is wrong with Sak?

Focus groups were also used to modify the vignettes prior to their incorporation into questionnaires to test the congruence among symptoms, labels, and action more widely.

In another study, different kinds of diarrhoea were identified in interviews through the use of a colour chart. This allowed the informants to distinguish "normal" and sick" stools on the basis of pigmentation as well as texture, frequency of emission. smell, and so on, with precision that was not possible through language alone due to hick of congruence between English and local colour terms (e.g., "yellow") [22]. The use of photographs might be explored in focus groups concerned with tropical diseases, for instance, to capture distinctive clinical signs and symptoms and to establish local ways of identifying and classifying diseases.

The success of the focus group depends largely on the skills of the moderator. Improvising the conduct of the discussion. managing aggressive and timid participants, fielding personal attacks, and deflecting conformity within the group (where appropriate) are not easy tasks. In addition, the moderator needs to be a skilful manager of group dynamics and to have a good grounding in the research topic and associated issues. Training is therefore imperative; four to five days were suggested prior to one focus group study [a15, 23].

Focus groups and informal group discussions

Conventionally, the term 'focus group'' is applied to a formally constituted, structured group that is brought together to address a specific issue within a fixed time frame and in accordance with clearly spelt out rules of procedure. While the literature does not address issues of validity that may he implicated by violations of these rules, or the constitution of the group per se, in natural field settings it is difficult and sometimes impossible to follow rules that were developed for application in industrialized. urban settings. We suggest that the focus group represents one extreme of a continuum of forms of group discussion, not dissimilar to the continuum of types of interviews, which range from informal conversations at one end to formal, structured interviews at the other. Each method may produce different data. but all provide data for analysis and interpretation by the social researcher. These are summarized in table 1.

Informal conversations occur whenever a researcher is in the field, walking around the village, stopping to chat with the people. On the move, it is possible to maintain single casual conversations and informal interviews of some duration. Settled on a house step or a mat, it is less likely: particularly during early stages of field work, when the researcher provides novelty to the villagers. a conversation of any length will attract others. In addition. depending on the cultural setting, domestic work may be undertaken in a group rather than on a solitary basis. again making one-to-

one researcher-informant interchanges difficult if not impossible. As a consequence. informal group discussions rather than dyadic conversations may be the norm. Where the field researcher plays an active role in directing the give and take, it becomes an informal focused group discussion.

Informal group discussions build on the social net works that operate in a natural setting. In a village. these typically include both kin and neighbours - a woman, her daughter, daughter-in-law, and unrelated immediate neighbours, for example. such natural clusterings represent, in a loose fashion, the resources on which any member of the group might draw, both in material terms and with respect to information and advice. This is a group that may weave or repair nets together. while watching over children and discussing the events of the previous day. It is precisely this natural social network that provides the script for managing an illness: what to do with a child with bloody diarrhoea, how to nurse high fever, or who to call in the case of threatened miscarriage. Decisions about such matters are rarely carried out by one care giver alone: people draw on those around them. As a result, such discussions provide fairly accurate data regarding the diagnosis and treatment of illness. choices of health services, and so on. In contrast to formal focus groups, however, they do not capture the whole range of attitudes and practices within a particular residential or cultural setting.

Unlike formally constituted focus groups, informal group discussions are neither time-bound nor limited in terms of participants. Participants can and do join and leave at will; the discussion may even begin and conclude as a one-to-one interview. Such a discussion can be made more formal, however, by reconvening the group to discuss a particular topic ("Could I come back tomorrow at the same time to talk more about skin disease?"), in much the same way that formal interviews are set up after informal discussions.

Informal group discussions, like informal conversations, take advantage of the natural setting. The researcher maximizes chance events and meetings to build up knowledge of the research topic. Informal discussions differ in many respects from the focus group as described in the research methods literature, but in natural field settings even the most formally organized focus group may be transformed into an informal group as uninvited villagers join or as original participants leave. Although, technically, established procedure calls for closure of discussion in the face of such disruptions, our inclination is to tolerate these exigencies and maximize any possible advantages that arise ("So should women be in the house when the husband comes back from the field?" "Isn't the clinic open at that time'?").

TABLE 1. Qualitative research methods

This is not to say that formal focus groups are not possible in the field. Although sessions conducted in a private house may present problems of space and privacy, participants themselves may take responsibility for ensuring group cohesion and maintaining the course of discussion. Chirawatkul (personal communication, 1990)) held focus groups that ran for around two hours, during which time the participants closed shutters to maintain privacy, and curious neighbours were deflected with the promise of a second discussion group to be convened the next day.

Moreover, the basic group structure is not always unfamiliar to villagers. Village decisions may be made by groups rather than individuals, and interventions based on community participation often use focus groups for health education, programme design. planning, and delivery [3, 24]. A nutrition education intervention program me in north-east Thailand. for example. designed to reduce the incidence of opisthorchiasis, used large discussion groups - age- but not sex-specific - to talk about the general issues. The participants were then divided into small age-specific groups to explore in greater detail the cultural and social issues associated with the consumption of raw fish (L. Manderson, field observations, 1990).

Where the integrity of a group may be broken to its disadvantage, either in terms of the flow of the discussion or the value of data collected, concurrent sessions may resolve the problem. A study of maternal management of diarrhoeal disease conducted concurrent focus groups for mothers and grandmothers to avoid imposing the values and practices of one age group on another [21].

Adapting the rules to the field

As indicated, focus groups do not always run as planned or as described in texts. Rules of organization, structure, and process are ideal but not always practical. A fixed time may work in industrialized settings, but not all people share the researcher's attitude to time.

It also held that, where the topic for discussion is culturally sensitive. the quality of information is enhanced through anonymity. While this may be easily realized in certain urban contexts, it may not be possible in small villages or in slum and squatter settlements. Moreover. in such settings, familiarity rather than anonymity may be the key to fluid discussions. Women, for example, may not be prepared to speak openly in front of strangers.

The sex, age, and other attributes of the moderator and note taker may also be critical in this respect. The use of tape recorders, though recommended, may interfere with the flow of conversation or the willingness of participants to join the group. In such circumstances, recording of the session may have to be suspended.

Focus groups are conventionally homogeneous; that is, participants are recruited for particular groups on the basis of some common characteristic to maximize interaction within the group and to avoid individuals dominating or withdrawing because of differences in class, sex. age, experience. acquired knowledge, or other characteristics salient to the discussion. However, homogeneity may not always be possible in the field, or its importance may need to be weighed against other key research objectives.

In some field settings it may be easier to attain homogeneity with men rather than women. A preselected woman may bring another woman with her; particularly if she is young, she may be accompanied by her mother-in-law or younger sister-in-law [16]. Where there are cultural restrictions on women's mobility, bringing them together from different villages to a central place may be still more difficult, even if transportation is provided. Interestingly, in some cases heterogeneity may mother-in-law be useful in assessing community attitudes and beliefs and in maintaining the flow of the discussion [16]. For example, difficulties in stimulating discussion on reproductive issues among younger women were resolved when an older woman who was present (the mother-in-law of one of the participants) started talking.

When homogeneity is the norm, the question arises as to how many focus groups are required to ensure adequate coverage and representation. It has been suggested that at least two focus groups should be conducted for each variable considered relevant to the topic [10, 13], but this can quickly become unmanageable. Sittitrai and Brown [15], for example, identify gender, age, socio-economic status, religion, sexual orientation, educational level. ethnicity, language group, urban or rural residence, and marital or coupling status as significant variables in research on sexual behaviour. If more than one area is included in the study, the number of focus groups multiplies accordingly. The cost of such a study can be prohibitive and the results cumbersome. Common sense decrees that some social categories, are, in fact, far more important than others, and that faithfulness to the principle of complete homogeneity in a small field setting would be impossible; that is, the entire village may have to be interviewed in order to ensure coverage. The alternative suggestion - that the target participants should be able to interact well and that focus groups should be conducted until no new information is gathered - is both a sensible and viable alternative [13].

When focus groups are used to complement other methods, the issue of how many focus groups should be conducted can be largely a matter of expediency. The question assumes more importance, however. when a study relies on focus groups as the sole method of data collection. One study used focus groups alone to assess community knowledge and perceptions and treatment patterns of diarrhoeal disease in seven cultural regions of Tunisia [23]. It also investigated differences in perception and practice among fathers, mothers, and grandmothers. Sixty-one focus groups were conducted over a period of four weeks during the peak diarrhoeal season. In total, 642 women were interviewed in the study by three moderators, who received four days' training prior to commencing research.

In contrast, another study conducted only one focus group with each of seven ethnic groups (with five or six participants per group) and an interview with one member of an eighth community [25]. The authors originally intended to conduct two focus groups per community, but this proved 'too difficult." The proceedings of each group were taped. transcribed, and summarized, and these summaries constitute the bulk of the monograph. The presentation of the material in the report implies that it is generalizable; that is, despite the obviously inadequate sample size, the authors have been rather cavalier in taking as representative the views and experiences of one group of women from one place. and in offering few caveats to their interpretation of the data.

There must be, obviously, a middle ground in all of this. Morgan [10] argues that large focus group studies generally are not practical. hut that the key to the size of a study (that is. number of focus groups conducted) is the variability of data required (ideas. experiences, opinions). Hence, he maintains that, once the point is reached at which the moderator can anticipate the outcome of a group discussion correctly before the group has met, no more focus groups need to be conducted. This point is normally reached, he maintains, after three to four groups:

Research that is exploratory in nature or simply aimed at "getting someone's perspective" will probably take only a few groups with relatively high degree of moderator imposed structure, but if the goal is detailed content analysis with relatively unstructured groups, then six to eight (or more) groups will be necessary.... But one group is never enough: you may be observing little more than the dynamics of that unique set of participants. [10]

Few economies are realized in doing a large number of groups, and the dilemma concerning the reliability and validity of data collected from purposively selected groups remains | 1()|.

Little in the literature suggests ways by which the existing shortcomings of focus groups could be overcome to enable them to be used more extensively as the sole or primary data-collection technique. Our preference is therefore a judicious mix of methods.

Since the aim of rapid assessment is to gain a general picture rather than to provide exhaustive documentation of a particular issue, focus groups may be both useful and efficient for collecting certain kinds of data. It should be stressed, however, that even studies that reportedly use only this method draw on other kinds of social information before beginning the study proper, to assess salient community categories (ethnicity, dialects), to identify culturally sensitive issues that may or may not be well explored in focus groups, and to gain vocabulary and a rough understanding of issues relevant to the development of focus group guidelines. Local researchers with prior knowledge of the cultural milieu and with indigenous language skills are obviously best positioned to undertake such studies.

The use of focus group research

As already indicated, focus groups. Iike other qualitative methods, can complement survey-based research in a number of ways (fig. I ). They provide a preliminary step in the development of a quantitative study. They generate ideas for narrowing the scope of the research, yield hypotheses for field testing. select appropriate wording of questions, and identify target groups for study. They may be used to explore and illuminate results of a quantitative study or to gain greater understanding about the reasons for certain trends. They are a means of gaining insights about perceptions, reported practices, population attributes, sources of knowledge, levels of awareness, opinions, attitudes, problems and fears. and the language (vocabularies) used to talk about all of these [31

Focus groups are used as a primary data-collection method, including for some topics that cannot be studied easily through quantitative methods or through individual interviews (e.g., sexual behaviour). Certain aspects of research on tropical diseases may best be explored in group discussions rather than individual interviews: attitudes toward the allocation of household resources for chemotherapy, the acceptability of bed nets to reduce the risk of malaria infection, or local terminologies applied to signs and symptoms of a disease. The full potential of their application awaits field investigation.

Focus groups have several advantages over both survey methods and other qualitative approaches [17, 26-28]. They appear to reduce the chance of questions being misunderstood by respondents and to reduce socially desirable answers (despite some problems of conformity within groups). The informal atmosphere of the group also should encourage participants to express views frankly and freely. On the other hand, the (usually) small sample and the purposive selection of participants does not enable quantification or allow generalization of the findings [3, 15].

Like other qualitative methods, the chance of introducing subjectivity in the interpretation of data - and in its collection - is high. and accuracy may be negatively affected if inadequate care is taken in preparing transcripts from the taped discussion. Moreover, the chance of introducing error in focus group research is particularly high when the interview is conducted in one language and then translated. While errors arising from translation are not unique to focus groups, the large quantity of data collected make checking by translation and back translation particularly awkward and time-consuming.

Methodological issues

In addition to these limitations, a number of methodological problems of focus group research have not been resolved. A major concern is the validity of conclusions where the quality of data is so dependent on the moderator, whose skills are central to the quality and quantity of the data obtained. We have already referred to the importance of training, but we also need to address issues relating to personal style. How much flexibility should a moderator have to encourage, cajole, push, and prompt the participants? To what extent do personal style and tone affect both participation and the validity and reliability of responses, or foreclose discussion because of disapproval or resistance?

Methods of analysis of focus group content also vary. A recent study conducted in collaboration with the International Labour Office [16] indicates that the validity of the findings depends largely on how the transcripts are analysed. In the study, seven sessions were both video- and audio-taped. Subsequently, the video tapes. transcripts, and study objectives were provided to five different professionals for their review and analysis. Preliminary findings suggest that drawing conclusions from the video tapes alone is problematic, and that, conversely, the use of transcripts is faster and easier. The findings also suggest that dependence on either the oral or visual record alone leads to some variation in interpretation and understanding. The study also indicated that, if content analysis of transcripts alone is done properly - that is, if care is taken to note which views are expressed, how many times they are expressed, and by how many participants - the conclusions drawn are fairly stable. Observations and notes of conversations further enhance the quality of the data.

For field researchers in poor countries. where despite best intentions basic equipment (for example, tape recorders) may break down and where other equipment such as video cameras may be unavailable or unacceptable to the participants, the accuracy of an analysis based on transcripts and tapes is good news. Time and cost are also key considerations. Our experience with transcription of both individual and group interviews supports the estimate that around five hours is required to transcribe one hour of recorded discussion [29].

FIG. 1. Integrating research methods

We have discussed in detail the variations in formality of group discussions and the number of groups needed to ensure 'reasonable coverage," because these are such obvious issues in the field. A related issue is that of group size, as well as the nature. advantages, and disadvantages of variations in the number interviewed: single-respondent interviews, dyads. mini-groups, and focus groups of different sizes. For example, one author observed that there was no correlation between group size and data collection, and that ideas produced in a group were not necessarily superior in quality to those produced in individual interviews [30]. Where group dynamics might be an important clue to a particular issue (e.g.. the acceptability of premarital sex). some kind of group interview might be valuable. But it is necessary to bear in mind that not all data can be collected in focus groups, and in some cases individual interviews may well be the most efficient and successful.

Formal or informal group discussions, focus groups, and individual interviews are not interchangeable - the issues implicated here need to be resolved in the field. While all gather information about attitudes. beliefs, and behaviour, they refer only to reported behaviour. actual behaviour requires observed cross-checking. As Morgan notes [10]. ". . . when there is a premium on the naturalistic ability to observe group behaviours and when the opportunity to observe such behaviour is readily available, some form of participant observation will be preferred over focus groups."

Further, given the importance of group dynamics, the kinds of responses that occur in group settings - these being culturally determined to a degree - need to be assessed accordingly. Experience in a focus group or individual interview may also affect a respondent's contributions, although it is not clear exactly how.

Conclusion

Focus group interviewing has, we believe, considerable potential as a rapid assessment procedure to enrich social and behavioural research. However, its methodological limitations need to be appreciated and its indiscriminate use discouraged: it is not a means by which the research process can be truncated [31]. The applicability of this method in tropical disease research needs to be tested in different field settings to assess not only the utility of group discussions for addressing particular questions but also whether these discussions can answer questions related to group dynamics in different cultural settings. Experimental studies are necessary to evaluate alternative qualitative approaches. in particular the merit of the focus group relative to individual or group interviews, and to assess the influences of procedural differences on the interpretation of research results. The focus group is perhaps best characterized, therefore, as a complementary tool which supports data gathered by other research methods.

References

1. Manderson L, Aaby P. Can rapid anthropological procedures be applied to tropical diseases? Health Policy Plan 1992;7(1):46-55.

2. Bogardus ES. The group interview. J Appl Sociol 1926;10:372-82.

3. Basch CE. Focus group interview: an underutilized research technique for improving theory and practice in health education. Health Ed Q 1987:14(4):411-48.

4. Merton RK, Kendall PL. The focused interview. Am J Sociol 1946;51:541-57.

5. Merton RK. Fiske M. Kendall PL. The focused interview. Glencoe. 111. USA: Free Press, 1956.

6. Goldman AK. The group depth interview. J Marketing 1962;26:61 -68.

7. Axelrod MD. Marketers get an eyeful when focus groups expose products, ideas, images, and copy etc. to consumers. Marketing News 1975;8:10-11.

8. Calder BJ. Focus groups and the nature of qualitative marketing research. J Market Res 1977;14:353-64.

9. Debus M. Handbook for excellence in focus group research: a special report of the HEALTHCOM project. Washington, DC: Porter/Novelli. 1988

10. Morgan DL. Focus groups as qualitative research. Newbury Park, Calif, USA: Sage, 1988.

11. Anker M. Focus group discussions. Geneva: World Health Organization, 1989.

12. Folch-Lyon E, Trost J. Conducting focus group sessions. Stud Fam Plan 1981;12(1):443-49.

13. Scrimshaw SCM, Hurtado E. Rapid assessment procedures for nutrition and primary health care: anthropological approaches to improving programme effectiveness. Los Angeles, Calif, USA: UCLA Latin American Center, 1987.

14. Long A. Scrimshaw SCM. Hurtado E. Epilepsy rapid assessment procedures (ERAP): rapid assessment procedures for the evaluation of epilepsy - specific beliefs, attitudes and behaviors. Landover, Md, USA: Epilepsy Foundation of America, 1988.

15. Sittitrai W, Brown T. Training manual on focus group discussions in human sexuality research. Bangkok: Chulalongkorn University, 1990.

16. Khan ME. Anker M, Patel BC, Hemlatha, Rao S. Use of focus groups in social and behavioural research: some methodological issues. World Health Stat Q (in press).

17. Knodel J, Pramulratana A. Focus group research as a means of demographic inquiry. Presented at the IUSSP Seminar on Micro-approaches to Demographic Research, Australian National University, Canberra, 3-7 Sept 1984.

18. Langer J. 12 keys to unlock qualitative research on sensitive subjects. Marketing News 1979:13:10.20.

19. Suyono H, Piet N, Stirling F, Ross 1. Family planning attitudes in urban Indonesia: findings from focus group research. Stud Fam Plan 1981;12(1):433-42.

20. McCracken G. The long interview. Newbury Park, Calif. USA: Sage. 1989.

21. Sani AbG. Punufimana AN, Seuseu NKF. Shawyer RJ. Use of clinical vignettes in rapid ethnographic assessment: a folk taxonomy of diarrhoea in northeast Thailand. Master's degree report. Tropical Health Program, University of Queensland, St. Lucia. Queensland. Australia, 1990.

22. Konare KF, Shaw DD, Jacobsen CA. Knowledge, attitudes and practices for planning oral rehydration strategies. Master's degree report. Tropical Health Program, University of Queensland, St. Lucia, Queensland, Australia, 1988.

23. Aubel J, Mansour M. Qualitative community health research: a Tunisian example. Health Policy Plan 1989:4(3):244-56.

24. Lewin K. Group decision and social change. In: Maccoby EE, Newcomb TM. Hartley EL. eds. Readings in social psychology. New York: Holt, Reinhart & Winston, 1958:197-211.

25. McCallum J, Gelfand DE. Ethnic women in the middle: a focus group study of daughters caring for older migrants in Australia. Report to the Commonwealth Department of Community Services and Health, Health and Community Services Research and Development Grants Committee. Canberra: National Center for Epidemiology and Population Health, 1990.

26. Folch-Lyon E, Macorra L. Shearer SB. Focus groups and survey research on family planning. Stud Fam Plan 1981;12(1):409-32.

27. Morgan DL, Spanish MT. Focus groups: a new tool for qualitative research. Qual Sociol 1983:7(3):253-70.

28. Freedman R. 'Fine contribution of social science research to population policy and family planning programme effectiveness. Stud Fam Plan 1987;18(2):57-82.

29. Zemke R, Kramlinger T. Figuring things out. Reading, Mass, USA: Addison-Wesley, 1985.

30. Fern EF. The use of focus groups for idea generation: the effects of group size, acquaintanceship and moderator on response quantity and quality. J Market Res 1982:19:1-13.

31. Merton RK. The focused interview and focus groups: continuities and discontinuities. Public Opinion Q 1987:51:550-56.

Contents - Previous - Next