Contents - Previous - Next

Appendix 8: Vitamin A content of foot

Analytical values for the vitamin A contents of food for a particular country or region can generally be found by contacting the national health authorities, or through the nutrition and/or agricultural offices in the Capital. Analytical techniques are updated from time to time and the most recent tables should be consulted.

Usually, in tables of food composition, vitamin A contents of food are expressed as IUs (International Units) or REs (Retinol Equivalents). The more recent tables use REs. In most tables, REs incorporate contents of retinol and carotene. In some food items, both of these forms are found, although animal food contains mostly retinol and plant food contains carotene.

Vitamin A contents are quite variable in different food items, and even within a particular food item grown in different parts of the world or exposed to different climatic handling/storage/preparation conditions. The highest levels of vitamin A are found in natural food items such as the livers of animals, carrots, red palm oil, and certain green vegetables and fruits. A recent publication entitled Food Sources of Vitamin A and Provitamin A was published in the UNU Food and Nutrition Bulletin 1992, 14(1):3-35. Another excellent source explaining the vitamin A levels of food is a publication from the International Vitamin A Consultative Group (IVACG) entitled Guidelines for the Development of a Simplified Dietary Assessment to Identify Groups at Risk for Inadequate Intake of Vitamin A. In 1993, C. West and E. Poortvliet released a publication for the Vitamin A Field Support Project called The Carotenoid Content of Foods with Special Reference to Developing Countries, which is a compilation of data on the vitamin A content of foods consumed around the world.

A general rule for vitamin A contents of food is to look at the color or to consider the part of the animal. Dark green vegetables, yellow and red fruits (excluding citrus) and vegetables, and red palm oil are rich sources of carotenoids. Animal sources include liver and organ meat, red meat, whole fish and fish oils, egg yolk, dairy products, and breastmilk. A range of values for vitamin A contents in food, together with a rating value to use in Module 5 tabulations is shown in the table below. While it is best to get a specific vitamin A value from a food composition source of original data, using an equivalent to guess at a rating is sufficient for the exercise in Module 5.

Keep in mind that substantial vitamin A can be added to the diet by eating small amounts of food high in vitamin A or larger amounts of food with modest levels of vitamin A. The bioavailability of vitamin A in food is also worth considering, and foods high in vitamin A must contain several properties for the vitamin A to be used by the body. Food must be palatable to the individual so that it is swallowed, and it must be digested; the diet must also have sufficient fat, protein, energy, and other key nutrients for dietary vitamin A to be utilized in the body.

Recommended levels of dietary vitamin A have been published by the Food and Agricultural Organization (FAO) of the United Nations in 1988, and are presented on the next page.

Examples of Food with Approximate Levels of Vitamin A and Rating Value to Use in Module 5

Food Example

Approximate Amt./Range RE/100g

Rating Value

Green leafy vegetables

3 -4
















coriander leaf



pumpkin leaf



drumstick; tree leaf






Root vegetables

1 -4




white potato



sweet potato, yellow








white maize



yellow maize


















10- 100


apricot, fresh



Animal foods


cow's milk



chicken's egg






beef liver, kidney



chicken liver








crevalle (Caranx sp.)



goby (Glossoqobium sp.)














coconut oil



seed oils, various

2- 100


red palm oil



narwhal blubber



fish oil, various



Breastmilk. human



NOTE. Values from Booth et al. (1992) and as reported in tables from regions where the manual was tested.

Rating Values to Calculate Vitamin A Contents of Diets

Rating Value

Amount of Vitamin A

Approximate REs






1 - 10



11 - 100







FAO Recommended Dietary Intakes of Vitamin A (RE)







1-6 yr



6- 15 yr










+ 100

+ 100




Appendix 9: Notes on selecting the field data-gathering team

As noted earlier in this manual, the field data-gathering team consists of at least three persons: the team leader and two research assistants. On the other hand, some groups who field tested these procedures found it advisable to have more research assistants, in order to avoid researcher fatigue and to expedite the completion of the data-gathering within the available time. With five, or even six, persons in the field work, the entire process outlined in the manual can be accomplished perhaps in four weeks instead of six. Decisions about the composition of the field team will of course depend on available funding, availability of suitable persons, and other factors.

Ideally, the team leader should be a person who has university training In some kind of community oriented social science. One obvious type of person for this task would be someone with a background in community nutrition, but many other types of persons would also be suitable. Obviously, the team leader should have a knowledge of food and good organizational skills, including the ability to direct and supervise the work of the assistants.

In the five sites where this manual has been used thus far, there have been two main types of team leaders. In some cases, such as Niger, the team leader was the principal investigator, a complete outsider to the local area, but with good fluency in two of the local languages (Hausa and French). She had extensive experience in the area of vitamin A programs, as well as other research in Niger. On the other hand, in the field study in Peru, the research supervisor was from Lima, but the team leader was from the local area, with a background in nursing. The assistants also had nursing training.

Criteria for selection of the assistants should put special emphasis on their familiarity with the local region, and its food culture, and socioeconomic system as well as their ability to establish good working relationships with people in the community to be studied.

The following selection criteria are to be considered when you put together the field data-gathering team:

i. All the team members should be persons who have good ability to develop friendly social relationships with the community people. Be especially wary of selecting persons who maintain social distance from villagers by use of more educated speaking style, manner of dress, and other symbols. In the same vein, team members should be persons who are non-judgmental concerning current cultural practices in the area. Thus, persons with healthcare and/or nutritional training and other service backgrounds, are usually willing and able to suspend judgment about food habits, hygienic practices, and other local behaviors during the course of this research.
ii. Of course all team members need to be available for the duration of the project. (Either persons who can take a leave of absence from their current duties or are currently unemployed.) If persons are selected who have other work and obligations, get a clear commitment concerning the numbers of hours and days per week the individual is available for data-gathering activities. iii. All members of the team need sufficient literacy levels so they can use the manual effectively and can write clear fieldnotes (see Appendix 5).
iv. Familiarity with the local language and culture is an especially important criterion, particularly in the case of the field assistants.
v. Persons with previous experience in community-based projects in the region would be likely to have better understanding at the outset, concerning the basics of data-gathering.
vi. Care should be taken that local persons are not seen as associated with a particular faction, especially political faction, within the community.
vii. The team members must be willing and able to visit all the different households in the study community. In some areas this can require walking in difficult, hilly terrain. In other communities there may be social difficulties for some people in going to households on the other side of the village.

In some areas you may find it very difficult to recruit educated persons for your data-gathering team. Here is an example from a rural area in Niger:

No college-educated persons were available, but the researcher, Lauren Blum, was able to hire one local woman with a high school education, and another who had not finished high school, but had a good level of literacy. The more educated woman was in her early thirties, an experienced mother, and fluent in French, Hausa, and Djerma, the three local languages. Her knowledge of the community and ability to develop good social relationships with all the people, more than offset her lack of special training in social sciences. The younger assistant, who was unmarried and childless, was somewhat less able to develop social relationships with the women in the household sample.

Appendix 10: Selecting representative samples

The information that you gather using this manual is intended to provide a balanced and fair representation of the target community and population. But your time and resources are short; usually you do not have the luxury of spending a lot of time enumerating every household and then generating a careful random sample. On the other hand, you will want to be sure that your observations, informal interviews, and your sample of household respondents are as representative as possible of the geographic and cultural subgroups and subdivisions of the community.

Representativeness: Age, Gender, Ethnic Groups, Geography

In any population there are different types of people, with different attitudes and information, and some of the differences are quite predictable. That is, you know from experience that information presented to you by males is likely to be different from that of females, and young people see the world differently from the senior generation.

You should try to have key-informants from different age groups and different neighborhoods or localities in your target area. Information about typical food use patterns should be gathered from persons who are current family food providers, but older persons who may be retired from cooking and food preparation, may be important sources of information concerning earlier food patterns, including former use of vitamin A-rich foods such as wild greens, leaves, etc.

Young persons, including small children, may be important informants concerning children's snacking and related food patterns, as well as their attitudes and taste preferences.

Ethnic groups and caste groups are usually different in their perspectives, behaviors, and knowledge, so try to include that dimension as well. In India, for example, villages very often have different subdivisions representing different caste groups. Thus, the data-gathering team should seek out key-informants in each of the different subdivisions or neighborhoods of the target village.

During the assessment you may not have time to be sure of representativeness of your key-informants, focus groups and all your miscellaneous sources of information. However, you should be constantly aware of the gaps in your information sources. Typically, you will be aware that you still have not talked with people from the other side of the tracks or, frequently enough, you will be aware that you have a serious imbalance of one gender or the other among your key-informants.

Use Maps and Diagrams to Chart Your Representativeness

As soon as you have a good working map of your target area, you can begin to use pins or other markers to identify the areas for which you have information. For example, you can use a pin for each key-informant, then inspect your map to see how you can increase the geographic representativeness of your information. Perhaps you will need to find some new key-informants in those areas that are still blank on your map.

Simple tables and charts can keep you reminded of the representativeness or balance of gender, relevant age groups, ethnic groups, and other differences in your client population. In some cases, you will be able to see from your charts and maps that you have information piling up in one area because all of your team members tend to go to the same area. Perhaps, after the first week or two, your team should disperse into different areas, or specialize in talking with different kinds of key-informants.

If your data-gathering team is all female and you find that you are not getting any interviews with males, then you will need to assign someone to get some interviews with males, to get a male-oriented view of food acquisition, crops, and food preferences.

Representative Sampling of Times and Places

We often think about representative sampling in terms of people. In fact, many books about sampling focus all their discussion on sampling from people, households, and other sampling units. On the other hand, it is important that you consider other types of representativeness. Observations of actual meals, including collection of 24-hour recalls, should be distributed as representative of the weekly cycle, for example. If there is more than one weekly market in the area, observations should be carried out in each of those sites, especially to note differences in the foods available, price differentials, and perhaps differences in the kinds of people who shop in the different locations. Similarly, if your project is in an urban location, different neighborhoods are likely to have different relationships to stores and other facilities.

Representative (Random) Sample of Respondent Households

In rare situations, you may find that there is a recent census of the target community, listing all the households With details of household composition. In such a case the drawing of a random sample can be quite straightforward. You would follow these steps:

i. Identify the list of all households with children between the ages of six months and six years.
ii. Assign a number to each household, starting with one.
iii. Select numbers from a table of random numbers. Each time one of those random numbers corresponds to a numbered household, that unit is added to the sample.
iv. For a proposed sample of thirty respondents, select fifty households, so you have a reserve to substitute for persons who are unavailable or unwilling to participate, as well as those who drop out due to illness or absence.

Drawing Household Numbers from a Hat (Instead of Random Numbers Table)

In most cases the total number of eligible households is not so large as to preclude your writing all the numbers on slips of paper and then drawing your sample in that time-honored folk method. The table of random numbers is then unnecessary.

Stratified Random Sample

In the majority of communities, there is some major division of the village or area-upper/lower caste, landowners/landless, uphill/lowlands, or central village/peripheries. In such cases it is wise to sample separately from the two sectors. That is, you would first prepare the two separate census lists (perhaps in consultation with key-informants) and then proceed to use the random numbers table.

Example of Random Sampling in Urban Location (Peru)

This example of sampling from a periurban neighborhood in Cajamarca, Peru, was done by Dr. Hilary Creed and her associates. The researchers had hoped to choose a sample of persons who had migrated to the urban area from the rural community they were also studying. However, they were unable to find families who were from that particular locality. Here is the procedure they adopted in order to get a representative sample in the community named San Vicente. The researchers developed their random sample in the following manner:

i. First, they divided the community into three sectors (because they had three field researchers assigned to this project).
ii. Within each sector the researcher selected an arbitrary starting point.
iii. The researcher then selected every fourth household from that starting point.
iv. If there was no one home, or the household did not meet the selection criteria (see below), the researcher then approached the next household and continued until a suitable and willing respondent was found.
v. After finding a suitable respondent, the researcher then moved on to the fourth household from that point and repeated the process.
vi. This same procedure, approaching the fourth house from the previous successful interview, was followed until the quota of respondents was complete.

Criteria for selection of respondents were as follows:

a. Must have a child between six months and six years of age.
b. Family is living m this community, not just visiting
c. Origin of the family in rural area was from a locality ecologically similar to the comparison community that was being studied at the same time.
d. Mother/caretaker willing to participate in the interview and to cooperate again when the interviewer returns for other parts of the structured interviews (Modules).

NOTE: In this form of selection (of every _______th household) the sample will be unrepresentative if there are considerable numbers of mothers regularly absent and those absences are due to the economic activities of the women. Key-informant interviewing will reveal the extent of this problem and can help in devising a supplemental sampling strategy to correct the bias.

For example, key-informants (and the respondents who are found at home) might report that the absentee women are engaged in wage work and/or market selling. In that case, an effort should be made to find enough of these women on weekends so that the sample will include adequate representation of women who work outside the household.

Contents - Previous - Next