Contents - Previous - Next
Option 1. Test the generalizability of the results and recommendations through additional ethnographic studies
Option 2. Test the generalizability of the results and recommendations through a structured interview format
Option 3. Test the generalizability of the results and recommendations through focus group discussions
Option 4. Develop and pretest messages and materials
Up to this point, you have collected information from a limited number of people. Interviews and other data collection exercises with key informants have provided an in-depth, insider's understanding of the local belief system as described by these local experts. In taking detailed case histories, you have expanded the sample to include caregivers of young children with diarrhea.
This next step will be determined by the objectives and scope of the research (see Part III 1), available resources, and other activities planned by the program or project. In most cases you will proceed to test whether the beliefs and practices described by the key informants can be generalized - that is, whether they are common to other members of the population. You may want to know if people in different communities or regions hold similar beliefs to those in your study area. You might start by identifying key informants in other areas and interviewing them using the ethnographic techniques presented in Part III.
Another way to determine to what extent the results of the ethnographic study can be generalized is by a structured interview. This consists of a limited number of questions that are posed to respondents in a uniform way. Your objective may be to determine whether other members of the study community hold beliefs that are similar to those of the key informants. You could do this by administering a structured interview to typical community members who were not selected because of any special knowledge or experience.
If the objective is to determine whether the results can be generalized to other communities or regions, it is also possible to develop or contribute to a more formal structured interview or survey. If a large household survey is planned by the program or project (or by an associated organization), it may be possible to contribute questions that will test whether the belief system you have described can be generalized.
Focus group discussions can also provide information about the generalizability of the study results. In these discussions, small groups of individuals with similar characteristics are presented with ideas or questions. The discussion is guided by a skilled facilitator.
Occasionally, when the intervention is to be limited to the small area in which the ethnographic research was conducted, the next step may be to develop and pretest messages and educational materials. While pretesting the materials with typical members of the community, you will indirectly learn whether they share the same beliefs as the key informants. In most cases, however, you should test whether the results can be generalized before proceeding with materials development.
Each of the
options, except for additional ethnographic data collection,
requires skills that are not described in this manual. Designing
and conducting surveys and developing and pretesting educational
materials requires collaboration with people who are experienced
in these fields. The following sections suggest how you and your
field workers can most effectively participate in these
activities. They do not attempt to provide a complete guide to
carrying out the activities.
You may choose to collect additional ethnographic data if the results of the first study suggest that there are subgroups in the community with a belief system that differs significantly from that of your key informants. Perhaps in taking detailed case histories you encountered a few caregivers who talked about different types of diarrhea or who expressed different ideas about how the body works. If some of the caregivers' responses do not agree with your understanding of the belief system, go back and ask them for more details. Perhaps they are not native to the community but come from a different ethnic group or another part of the country. If they are unable or unwilling to give you the detailed information you need, ask them to whom they go for advice when their children are ill. This may help you find others to explain the alternative belief system.
One way to test whether the belief system you identified in the study community can be generalized to other members of the population is to start from the beginning of Part III and conduct another ethnographic study in a different location or among a different religious or ethnic group. If you find that the terminology and beliefs expressed by the new study community are very different from those of the original community, you should gather the same amount of information (initially interviewing seven to 10 key informants and taking about 20 case histories).
country is very large or has groups of people that are
geographically very isolated from each other, it is more common
to find that general beliefs about illness and physiology are
very similar, although the words used to describe them are
different. In that case, you may be able to decrease the amount
of data collection, perhaps interviewing only four to six key
informants and taking 12 to 15 case histories.
1. Select questions for the structured interview
2. Plan and conduct a structured interview
In a structured interview, all respondents are asked the same set of questions in the same way. The results can be quantified and presented in numerical form. The strength of this method is that it permits you to estimate how common certain beliefs are, that is, what percentage of respondents express them.
Structured interviews can be more tiring than unstructured interviews, both to the interviewer and to the respondent, because they do not follow the natural flow of conversation. Therefore they must be brief, requiring a maximum of 1 hour per respondent.
this time limitation, structured interviews cannot ask about all
aspects of the belief system and cannot probe for further
information about responses. Therefore, when designing questions
to test the generalizability of ethnographic findings, it is
important to focus on a limited number of key questions. The
ethnographic research team's most important role in designing a
structured interview is to guide the selection of key questions
and the phrasing of questions in a culturally appropriate way.
Although one question may serve more than one purpose, each question should be carefully chosen to meet at least one of the following objectives:
To determine whether the beliefs described by key informants are common among other members of the community or the population;
To explore the appropriateness of possible approaches and images to promote AHM;
To obtain background information about the respondent.
Questions to test whether beliefs are common
These questions test whether certain beliefs are shared among members of the population or whether there are subgroups that hold different beliefs. You should focus on beliefs that are particularly relevant (either helpful or harmful) to promoting AHM of diarrhea and that form the basis of possible approaches for promoting AHM.
EXAMPLE: In Baluchistan, participants reported that "overeating" was a common cause of diarrhea in children and that children should always be kept "a little bit hungry" to prevent diarrhea. One of the structured interview questions was:
"If a child is taking powdered milk or solid food, should he be allowed to eat as much as he wants or not? Why/why not?"
EXAMPLE: As part of the structured interview, interviewers determined whether any child in the household had had a recent episode of diarrhea. If so, the mother was questioned about associated signs and symptoms and asked, "What kind of diarrhea do you think it was?" These questions checked the relationship between signs and symptoms and types. They also confirmed that the respondents identified the same types of diarrhea as the key informants did.
EXAMPLE: Researchers in Indonesia wanted to know whether a belief about kuman invisible living creatures that cause illness, was common in the population. Respondents to the structured interview were therefore asked, "Have you ever heard of kuman? " If the answer was yes, they were asked "What: are they?" If the answer was no, they were asked if they had ever heard of "invisible creatures that cause illness," to determine whether there was another word for the same concept.
Questions to explore approaches and images
You will want to ask some questions about each possible approach to determine which might be most effective in the community.
EXAMPLE: In Baluchistan the study team considered promoting ORS as a liquid to prevent and treat thirst. They asked:
"What type or types of diarrhea make children thirsty?"
"Why do children get thirsty during diarrhea! episodes?
If the respondent's child had a recent episode of diarrhea: "Did you give anything for thirst?"
If the answer was yes: "What?" "How much?"
They also considered promoting ORS as a liquid that replaces water lost during diarrhea. They asked:
"Where does the water in diarrhea come from?"
"Is water loss dangerous?"
If the answer was yes: "What can happen?"
"Do you know anything that can be done about water loss? "
Questions to identify respondents
Be sure you have a way to identify each household, either by the name of the head of the household or by house numbers. For each respondent you will probably want to record age, educational level, ethnic group or caste, number of children, village or neighborhood of residence, and some indication of socioeconomic status. This will ensure that your sample includes respondents of different ages, educational levels, ethnicities, etc.
identifying information may be useful in recognizing subgroups
with different beliefs. Perhaps younger caregivers or those who
are literate hold different beliefs.
This section describes some of the tasks and issues involved in planning and conducting a structured interview. If you have no training or experience in survey research and plan either to conduct a structured interview without expert assistance, or to work with a team to conduct a large survey, you should consult the references at the end of this section.
Train the field workers
Structured interviews can provide quantitative information about what percentage of people interviewed express a specific belief or report a specific behavior. Structured interviews differ from unstructured interviews in that the procedures for selecting respondents and the order and form of the questions are defined in advance. Field workers who have been working with ethnographic methods usually need retraining in these formal questioning techniques.
Pretest the questions
A structured interview questionnaire usually requires several rounds of pretesting. Drafts of the questionnaire are administered to a small number of people who are similar to those in the study population but who will not be part of the actual study. Pretesting is necessary to: 1) ensure that respondents understand and interpret the questions as you intend them to, 2) estimate the kinds and ranges of response you can expect to each question, 3) determine how long it will take to complete the questionnaire, and 4) format the questionnaire so that it is easy to administer and allows enough room to record responses.
Precode the questionnaire
The codes for entering the information into the computer should be included on the questionnaire. This will save a tremendous amount of time and reduce the error rate in translating written data into computer codes.
The households you select for interviewing should represent the subgroups of the community. If you simply ask your key informants and local leaders to suggest caregivers to interview, you will most likely be referred to the friendliest, most verbal, and best-educated people in the community. However, you should collect information from all types of people in the community, including those who are more isolated and less educated Researchers who conduct structured interviews usually use a sampling method (a way of choosing respondents) that ensures that every household in the community has an equal chance of being selected. A variety of sampling methods are described in the references at the end of this section.
In some cultures and settings, several caregivers of young children may reside in the same household. Interviewers should follow a uniform procedure for selecting among possible respondents within a household.
Supervise the data collection
interviewers who are conducting structured interviews must ask
the questions in the same way and use the same criteria and
format for recording responses. Although uniform procedures can
be taught during training, it is important to supervise the
interviewers intermittently during data collection to be sure
they continue to follow the established guidelines. Checking
questionnaires shortly after completion will help to identify
missed questions or unclear responses so that the interviewer can
return to the respondent for clarification.
Focus group methods are not presented in this manual, nor were they used during either of the field tests. We recognize, however, that many researchers have extensive experience conducting focus groups and will wish to use them to test the generalizability of results and recommendations.
excellent references on these methods are listed at the end of
1. Draft scripts to present the approaches
2. Present the scripts to small groups of caregivers
3. Develop illustrations for the scripts
step the possible approaches to promote AHM are presented to
people in the community to see if they make sense and if people
like them. We recommend that you first present the approaches
verbally only. This will allow reactions to the idea of the
approach to be separated from reactions to the pictures used to
In Part III: 6.5 you noted the messages that you wanted to communicate through the two or three best approaches and listed the images that might be used to explain the approaches. Now decide how you would present and explain the message to a caregiver of a young child. The following section continues the example of Baluchistan. The approach is "ORS goes into the veins and puts back the water lost during diarrhea." The image of a tree planted in a hole was chosen to help explain the approach. The following script was written as a test:
ORS is a special fluid to give during diarrhea. It goes from the stomach directly into the veins and puts back the water that is lost during diarrhea.
To understand how ORS works, compare the veins of the body with the roots of a tree. In places that are very hot and dry, people usually plant trees in holes to help the water go into the roots.
The roots of the tree are like the veins of the body. Just as the roots help bring water and food to the rest of the tree, the veins help bring water and food to the rest of the body.
When a child has diarrhea, it is important to help water go back into the veins. A little salt and sugar mixed in water will help do this. Salt and sugar work for the child as the hole works for the tree. They help water go into the veins just as the hole helps water go into the roots.
That is why the ORS package has salt and sugar in it.
When you put water in the hole, not all of it goes into the roots. Some may go into the ground and some may go into the grass. Animals may drink some of it. Still, a lot of the water goes where it is needed - into the roots.
In the same way, when you give children ORS, some, but not all, of it goes into the veins. The child may vomit some and some may come out with the diarrhea, but this is okay. A lot still goes where it is needed - into the veins. That is why you should keep giving ORS even if the child vomits or if some of it comes out in the diarrhea.
The size of the hole in which the tree is planted is important. If the hole is too small, water will spill out and not very much will go into the roots. If the hole is too big, the water will spread out and be wasted. Very little will go into the roots.
In the same way, the amount of salt and sugar in the water is important. If there is too much or too little, the ORS will not work in the right way. That is why it is important to measure the water and mix the solution according to the instructions.
After you have drafted the scripts for the two or three most useful approaches, you will be ready to present them to small groups of caregivers. Try to limit the groups to two or three members. Begin by explaining that you are trying to develop messages about how to care for children with diarrhea, and that you are going to read different messages. You would like them to tell you which of the messages they like better.
Read the first message slowly and allow time for the group to ask questions. Make note of any questions that are asked. Read the second one slowly, again allowing time for questions. Ask the group which message they liked better and why. Make note of all comments.
Present only two scripts to each group, even if you are testing more than two. Be sure to record any misunderstandings or misinterpretations of the messages. Do not give the impression that you are disappointed with the answers given or that people gave "wrong" answers. This would discourage further discussion.
Ask the group to recall as much as possible about the first message, recording all the points that are mentioned. Do the same for the second message.
Ask the group how the message could be improved. Would they suggest different examples? Different wording?
process with different groups until each message has been
presented at least five times. Change the order of the scripts,
sometimes presenting script A first, sometimes script B. If you
are testing more than two scripts, present them in different
You may want to work with communication specialists to develop illustrations of the approaches. Community members may find it difficult to respond to scripts that are read to them, and presenting illustrations along with the text may help their understanding and increase interest. These preliminary illustrations may serve as the basis for communication materials that are developed later.
Usually simple sketches are best. For approaches that contain several messages (as in the tree-in-a-hole metaphor from Baluchistan), several illustrations will be needed. Later these can be organized into a variety of formats, including counseling cards, comic books, flipcharts, or videos.
Some of the steps in developing illustrated health education materials are outlined below. Becoming familiar with the process will help you work with people who are experienced in this area. Details are presented in the references.
Summarize the messages and ideas for communicating them
Work with one approach at a time. Make a worksheet with three columns. In the first column, write down the messages to be communicated through the approach (see Part 111.6). In the second column, put down ideas for pictures to illustrate the messages (for example, a weak-looking child, a tree planted in a hole, a leaky oil lamp). Use the third column to record ideas for text to accompany each illustration.
Draft the design and layout
Using crude stick figure drawings, arrange the illustrations and text in a format that can be presented for pretesting. Usually each page will express one message by one illustration and one or two lines of text.
Work closely with an artist
Give the artist a copy of the layout and explain the approach and the individual messages that you wish to convey. Providing this context will help the artist understand the desired emphasis in each illustration and work more efficiently.
Prepare pretesting forms
Standard forms are used during pretesting to ensure that all information is collected in an organized way. Usually three forms are needed:
Background form for information about the people who reviewed the materials;
Data form for individuals' reactions to the illustrations and text;
Summary form to quickly present information in the data forms.
Select respondents for pretesting
For formal pretesting, usually the interviewer should work with only one individual at a time. Respondents should represent the audience for which the materials are intended. It is usually best to pretest materials with people you have not interviewed before.
Conduct the pretest
A pretest tests the ideas and the pictures, not the respondent. First the respondent is shown the picture and asked, "What do you see?" Then the interviewer reads or shows the text to the respondent and asks, "What do the words mean to you?" Next the respondent is asked, "How do you feel about the pictures and the words?" and is encouraged to suggest improvements to them. Responses are recorded on the pretest data form.
Summarize the results
Using criteria that have been determined in advance, the pretesting team assesses each respondent's reaction to each picture and text as either "OK" or "not OK." The results from all respondents are summarized to determine whether the individual illustrations and accompanying text are acceptable, or if they need revision and further pretesting.
Revise the materials
Materials are revised according to the recommendations, and preparations are made for another round of pretesting. The entire process is repeated until pretesting confirms that most respondents like and understand the messages as expressed by the illustrations and text.
1 Refer to the references at the end of this section for further suggestions on testing and developing messages with groups.
Fink, A., and Kosecoff, J. 1985. How to Conduct Surveys. Sage Publications, 2455 Teller Road, Newbury Park, CA 91320
A step-by-step guide to survey design and administration. It has a particularly strong section on how to avoid problems with the wording of questions. Unfortunately, the examples refer to educational research in developed countries. The principles, however, are clearly explained and are easily applied to surveys conducted in developing countries.
Nichter, M., and Nichter, M. Health education by appropriate analogy: Using the familiar to explain the new. Convergence 1986;19(1):63-72.
This article discusses the use of analogy as an alternative means for health education. It describes the steps in framing appropriate analogies for health messages and in developing the analogical messages through dialogue with community members. The examples refer mainly to nutrition and family planning, but the process can also be applied to presenting and developing approaches to promote AHM of diarrhea.
Rasmuson, M., Seidel, R.E., Smith, W.A., and Booth, E.M. 1988. Communications for Child Survival, HealthCom, Academy for Educational Development, 1255 23rd Street, NW, Washington, DC 20037
An excellent resource document that provides an overview of public health communications and methods and includes case studies, detailed examples, and a bibliography.
Zimmerman, M., Newton, N., Frumin, L., and Wittet, S. 1989. Developing Health and Family Planning Print Materials for Low-Literate Audiences: A Guide. Program for Appropriate Technology in Health (PATH), Communication Department, 1990 M Street, NW, Suite 700, Washington, DC 20036
An excellent step-by-step guide to developing and pretesting messages (both text and pictures) for health projects in developing countries. It is available free of charge.
Contents - Previous - Next