3.2 Contents of a survey

The contents of a nutrition survey are arranged according to the objectives as described in Chapter 1.2. In particular, if an objective of a project is to improve the nutritional status and living conditions, it is necessary to identify the causes of the nutritional problems in the population. Figure 5 shows the primary factors causing nutritional problems, all of which are based on secondary causes. It can be observed from the problem tree that a low nutrient intake (01) and/or a high prevalence of infectious and carrier-borne diseases (02) are responsible for a poor nutritional status (0). Similarly, both causes result from inadequate availability of food at the household level (1), poor social relationship and caring capacity (2), inadequate health services (3) and unfavorable environmental conditions (4).

Figure 5. Causal model of malnutrition
s53.gif (7934 Byte)

Generally, the probability of a cause being responsible for a poor nutritional condition becomes smaller as one progresses more deeply into the levels of causes, and as one moves further away from the core problem. Consequently, it makes little sense merely to follow one branch of the tree down to the details (e.g., supply of nutrients at the household level), if other branches remain neglected (e.g., health services or environmental conditions).
The content of the baseline survey is determined by the theoretical cause-effect-relationships of the problem tree. Care must be taken to ensure that all problems are considered before going on to the next lower level of causes.

This chapter will examine in detail the content of a nutrition survey. Nevertheless, these are only guidelines, as the potential variables have to be examined for their relevance and necessity for each location.

The wording of the sample questions is merely one possible way to obtain the data needed for each of the survey variables. Obviously the development of questionnaires, the wording of questions and the selection of the answers must consider socioeconomic and cultural realities and should sound natural in the language used. Local specialists and project staff can be of great assistance in this area.

The variables presented here are divided into two groups:

1.) Variables 1 - 40 (sub-chapters 3.2.1 - 3.2.8) that are related to the household level.

2.) Variables 41 - 82 (sub-chapters 3.2.9 - 3.2.12) that are related to the individual level.

3.2.1 Information on survey organization

Each questionnaire should begin with a cover page that helps to relate the questionnaire to the survey, the enumerators and the household.

1.) Household number

The household number should be filled into the questionnaire by the survey leader already before the forms are distributed to the survey teams as they leave to the households. This ensures that all visited household have only one unique household number.
Variable code: HOUSEHNO

2.) Survey team number

The survey team number should be filled into the questionnaire by the survey team at the begin of the survey before visiting the households.
Variable code: SURVTNO

3.) Supervisor number

In case of larger surveys in which the survey teams are coached by supervisors, the number of the corresponding supervisor should be filled in.
Variable code: SUPERVNO

4.) Date of survey

The date of survey should be filled in by the survey team before starting the data collection in the households. The date of the survey is needed to calculate the age of the surveyed children.
Variable code: SURVDATE

5.) Location of household (village, suburb)

The code of the village or suburb should be filled in before the survey team visits the households.
Variable code: HSHLOCAT

3.2.2 Demographic data of household

The demographic situation of a household is documented in a table. The table is used as an information source for several items. Once all relevant data have been extracted from the table, the page with the table must be destroyed to guarantee the anonymity of the respondent.

The name of each person who belongs to and is found in the household is written on the table. Although the names will not be used afterwards, providing each name helps remind the interviewee how many persons live in the household. Based on the name list, the number of persons living in the household will be determined later.

In various cultural situations the number of family members is not constant and the definition of who is attached to the household is different from our way of thinking. Temporary visitors in the household staying with the family may be considered family members. Also the father may work outside the home for several weeks at a time, and/or the children may leave home when very young. Furthermore, in some societies children who are not yet baptized would not be counted.

Definition: A household includes all persons:

The second part of this definition may not be applicable in rural areas, as in agricultural societies other household arrangements may exist.

The age of different household members may contribute important information about possible causes of malnutrition. In particular, infants from mothers who are younger than 16 or older than 40 have a higher risk of low birth weight. Furthermore, birth frequencies less than two years are also associated with a higher risk of low birth weight.

The age of children, and also of adults, is often unknown (see section In such societies, a suitable way of asking this question must be found, or the question must be left unanswered for all household members except the age of the mother and the children under five years.

The question whether the household member has earned money during the last three months applies particularly in urban areas, as in rural areas most family members take part in agricultural work and it is difficult to distinguish between working and non-working members. However, this question could be relevant for landless rural families.
Stay in the household
Earning money

Family status:
1) Father    4) Grand parents   7) Others
2) Mother    5) Aunt/uncle      9) No answer
3) Children

1) Male      2) Female

Stay in household:
1) Stays permanently
2) Does not stay permanently, but >8 months in the household
3) Does not stay permanently, but 4-8 months in the household
4) Does not stay permanently, but <4 months in the household
5) Moved away but sends money
8) Don't know
9) No answer

Earning money:
1) Earned no money in the last 3 months
2) Earned no regular income in the last 3 months
3) Earned regular cash income in the last 3 months
8) Don't know
9) No answer

6.) Observation: How many persons live in the household?

Variable code: HSHMEMNO
The information is calculated from data in the household table.

7.) Observation: How many children under 5 years live in the household?

Variable code: NOCLDMEM
The information is calculated from data in the household table.

8.) Observation: Child number

Variable code: CHILDNO
The household table indicates the children under five years from whom the data for the individual questionnaire will be collected. In case that a household has more than one child younger than five years and information from all of these children shall be collected, the number of individual questionnaires should be attached to the household questionnaire. Each child of this age group should be indicated with a number starting from number 1. The numbers are then filled into the attached individual questionnaires corresponding to each child to be surveyed.

9.) Observation: Is the head of the household male or female?

1) Male   2) Female   9) No information
Variable code: HSHHDSEX
The information is collected from the household table.

10.) Observation: How old is the mother who lives in the household?

Variable code: AGEMOTHE
The answer is given in years and is collected from the household table.

11.) Observation: How old was the mother at the birth of the oldest child?

Variable code: AGEOLDCH
The information is calculated from data in the household table.

12.) Observation: How old was the mother at the birth of the youngest child?

Variable code: AGEYOUNG
The information is calculated from data in the household table.

13.) Observation: How many cases exist with birth spacing less than 2 years?

Variable code: BIRTHSPA
The information is calculated from data in the household table.

14.) Observation: How many household members earned money in the last 3 months?

Variable code: HSHMMNO
The information is collected from the household table.

15.) Observation: Has the mother earned money during the last 3 months?

1) Yes    2) No    9) No answer
Variable code: MOTHMONE
The information is collected from the household table.

3.2.3 Ranking of problems by the observed household

One function of the baseline survey is to assist in the identification of appropriate interventions to solve the problems faced by the target group. Consequently, the baseline survey should obtain a complete (holistic) description of reality. In addition to scientifically based observations and investigations, the perceptions of the target group are an important part of this reality.
The sustainability of the results of an intervention program is drastically reduced if the opinion of target groups is not included in the diagnosis of the living conditions and nutritional situation and in planning the intervention. For this reason, it is essential that the survey reflects the viewpoint of the community.

Questions concerning nutritional and health problems should gradually move from general aspects to specifics, thus enabling the ranking of different nutritional problems to be explored. This technique also prevents the interviewer from assuming that the interviewee has a particular nutritional problem.

The following problems can only function as examples and have to be changed according to each region. This requires an accurate knowledge of the local area through study and reports on the experiences of others.

No generally applicable rule can be given for obtaining a general assessment of the living situation during the interview. However, it has often been shown that this question should be asked later in the interview, because at the beginning of an interview the interviewee is often reluctant to express an opinion on these matters.

16.) Question: Which are the major problems in your daily life?

The respondent shall identify up to three main problems.

In rural areas:
Problems (not more than 3 items) Answers
1) No or little land
2) Low yield from the land 
3) Widely scattered fields 
4) Little income 
5) Poor living conditions 
6) Disputes with neighbors 
7) Educational problems for the children 
8) Frequently ill 
9) Little food 
10) Poor or inadequate water supply 
11) Inadequate energy supply (wood, electricity, etc.) 
12) Other problems 
13) No problems 
1) Yes     2) No

In urban areas:
Problems (not more than 3 items) Answers
1) No or unclear right of residence 
2) Unemployment or under-employment 
3) Much time spent on getting to work 
4) Little income 
5) Poor living conditions 
6) Disputes with neighbors 
7) Educational problems for the children 
8) Frequently ill 
9) Little food 
10) Poor or inadequate water supply 
11) Inadequate energy supply (wood, electricity, etc.) 
12) Other problems 
13) No problems 
1) Yes   2) No

Before conducting the survey the relevant answers that can be expected in the survey area must be identified.

Variable codes:

3.2.4 Socioeconomic household data

17.) Question: Who is mainly responsible for bringing up the children?

1) Mother  5) Grandmother/grandfather
2) Father  6) Other relative
3) Both, mother and father  7) Other non-relative
4) Sister/brother  9) No answer
Variable code: CARERESP

18.) Question: What occupation has the head of the household mainly engaged in during the last three months?

1) No occupation       8) Dealer or trader
2) On daily wages      9) Salesperson
3) Domestic servant   10) Civil servant
4) Industrial worker  11) Military or police
5) Farmer             77) Other
6) Fisherman          99) No answer
7) Craftsperson
Variable code: OCCUPACI

In urban areas, there are often strict distinctions between occupations. If the head of the household has more than one occupation, the occupation category with which he/she most identifies should be entered into the questionnaire.

19.) Question: Were you born in this town/city or from which part of the country did you come?

Variable code: ORIGIN
This question is intended primarily for urban areas, where large ethnic and cultural differences in the community can exist due to a high migration rate. The question should also be asked in rural areas if there is a high migration rate from other areas in the country. Before conducting the survey, the possible answers must be identified.

20.) Question: To which ethnic (or religious) community do you belong?

Variable code: ETHNIREL
Before conducting the survey the possible answers must be identified

21.) Question: How much formal education have the parents had?

1) < 3 years of schooling (illiterate )
2) 3 - 5 years of schooling (can read and write)
3) 5 - 11 years (attended secondary school)
4) Higher education (> 11 years)
Father              Mother

The UNESCO definition stipulates that a minimum of 3 years of uninterrupted schooling is required for a person to achieve a sustainable level of reading and writing ability.

22.) Question or observation: The walls of the room in which the children sleep are constructed of what material?

1) Timber   5) Pasteboard
2) Mud      6) Cement/concrete
3) Stone    7) Other
4) Brick    9) No answer
Variable code: WALL
The decision must be made before the survey whether this will be a question or an observation.

23.) Question or observation: How many bedrooms are there in the house or residence?

Variable code: NBEDROOM

24.) Calculation: How many persons on average sleep in one bedroom?

Variable code: PERSOBED
Total household members / bedroom
This variable shall not appear in the questionnaire.

25.) Question: From where was the household drinking water obtained yesterday?

1) Own public water supply  5) Well/spring
2) Public supply outside    6) River/canal
3) Water traders            7) Other
4) Rain water               9) No answer
Variable code: WATER

26.) Question: Are there especially critical months when water is scare?

     In which month does it start?
     In which month does it end?

0) No supply problem  5) May         10) October
1) January            6) June        11) November
2) February           7) July        12) December
3) March              8) August      88) Don't know
4) April              9) September   99) No answer

This information is used to fill in the chart below.
Month Answer

Answer: 1) Yes   2) No


27.) Question: How was sewage/human waste disposed of yesterday?

1) Connection to a public 5) River/lake
   sewage system          6) Sea
2) Latrine                7) Other
3) Garden                 8) Don't know
4) Field/woods            9) No answer
Variable code: SEWAGE

28.) Question or observation: How was garbage/household waste disposed of yesterday?

1) Public garbage collection   2) Disposed of openly (street,..)
3) Burned                      4) Buried
7) Other                       9) No answer
Variable code: GARBAGE

29.) Question or observation: What form of energy was used for cooking yesterday?

1) Electricity            2) Wood, bought
3) Wood free collection   4) Wood from own property
5) Petroleum              6) Dried animal manure
7) Other                  9) No answer
Variable code: ENERCOOK

30.) Question or observation: Does the house have electricity?

1) Yes, directly from the public supply
2) Yes, via a neighbor
3) No
7) Other
9) No answer
Variable code: ELECTRIC

The following questions are applicable in rural areas.

31.) Question: How large is your agricultural setting?

Variable code: FARMAREA
Size in hectares or local units. Experience shows that this question is not always accurately answered, as farmers are often apprehensive that the answers will be used by government offices for other purposes (e.g., tax assessment). In agricultural operations where a farmer raises livestock on his own land, the size of the pasturing land should also be asked.

32.) Question: What is the ownership status of the land?

1) Own land                7) Other
2) Leased land             8) Don't know
3) Owned and leased land   9) No answer
4) Public land
Variable code: LANDOWN

33.) Question: Which are the most important crops you cultivate?

The table is completed to collect information on crop cultivation. Questions are asked on the four most important staple crops (e.g., wheat, rice, maize, potato) and their predominant use. The pilot study is used to identify the four crops and before the survey begins, the names and code numbers of the crops are entered in the first column. In the last column the number code of the predominant use is entered.
Cultivations Use


1) For own consumption
2) For sale
3) Both of approximately equal importance
4) Not cultivated
9) No answer
Variable code: CROP1, CROP2, CROP3, CROP4
Each of the four surveyed main crops needs its own variable code. It is recommended to call them:

34.) Question: Which are the most important types of livestock that you raise?

A table is completed while obtaining data on raising livestock. Questions are asked concerning the four most important species of animals (e.g., buffalo, pig, goat) and their predominant uses. The pilot study is used to identify the animal species, and before the survey begins the animal species and code numbers are entered in the first column. The response code of the predominant use is then entered in the last column.
Animal species  Use


1) For own consumption
2) For sale
3) Both of approximately equal importance
4) Not raised
9) No answer
Each of the four surveyed main livestock needs its own variable code. It is recommended to call them:
If there are other important agricultural products in the region (i.e., fruits, vegetables, trees, shrubs, fisheries etc) additional questions should enquire about the importance of these for the household.

3.2.5 Dietary pattern

Nutrient intake can be studied using several different methods; two commonly used ones are the weighing method and the 24 hour recall method. In the weighing method, dietary intake is determined by calculating the net amount of food and drink consumed for each person over a 24-hour period (the difference between the weight of the items offered to the individual and the weight of the items left over). In the 24-hour recall method, specially trained interviewers ascertain which foods and drinks have been consumed in the last 24 hours and the amount of each item consumed.

The nutrient intake from the food and drink consumed can then be calculated using nutrient value tables or more commonly now by using computer software packages. If the nutrient intake is compared to nutrient requirements (see chapter 6.4), it is then possible to hypothesize about the probable causes of nutritional problems. In the Nutrition Baseline software optionally a quantitative Food Recall is integrated which can also be used for the weighing method and the 24 hour recall. If the mean food intake for the mother and/or the child is entered the program calculates automatically the nutrient intake and the nutrient density for all nutrients which are selected in the main program.

Unfortunately, the weighing method and the 24-hour recall method are very time consuming and extensive training of the survey personnel is necessary. Also, in order to obtain representative data, many days of observations are required, particularly for some important nutrients (e.g., vitamin A). Additionally, the reproducibility of the data can also be affected in other ways, such as seasonal fluctuation in the availability of foods.

After considering the advantages and disadvantages, neither the weighing method nor the 24-hour recall method should be employed in this type of nutrition survey. It is much more important that the frequency of nutrient intake be determined. This method does not allow to obtain an exact picture of the nutrient intake, but it is possible to derive qualitative statements on the potential nutritional problems within a household.

The nutrient intake in the family can provide insight into the usual nutritional practices of the target community. This information is important if supplementary food or nutrients are necessary so that usual nutritional practices are considered as far as possible.

35.) Question: Which foods are eaten in your household?
Food sources Frequency of
Staples (grains, tubers, legumes):
Mixed or processed foods 
Meat (large livestock, e.g. cow, pig, sheep) 
Meat (small animals, e.g. poultry, rabbit) 
Fish, seafood 
Egg, egg products 
Milk, milk products 
Plant fats 
Animal fats 
Fresh green leafy vegetables 
Other fresh vegetables 

0) Never
1) Strong seasonal fluctuations
2) Monthly, once
3) Monthly, more than once
4) Weekly, once
5) Weekly, more than once
6) Daily, once
7) Daily, more than once

Food items: Variable codes:

Starches (grain, tubers, legumes) STAPLE1, STAPLE2, STAPLE3, STAPLE4
Pre-prepared food PREPARED, SNACKS
Energy foods (fat, oil, sugar) OIL, PLANTFAT, ANIMALFT, SUGAR
Fruits and vegetables GREENVEG, OTHERVEG, FRUITS

The following questions provide information on the seasonal nature of food supplies and nutritional status. The nutritional condition of rural communities can vary widely during a year, especially if agricultural production is unevenly distributed for climatic reasons (e.g., large variations in rainfall distribution) and if storing food or earning supplementary income is not possible.

36.) Question: Are there especially critical months for food supplies?

In which month does it start?
In which month does it end?

00) No supply problem  05) May         10) October
01) January            06) June        11) November
02) February           07) July        12) December
03) March              08) August      88) Don't know
04) April              09) September   99) No answer

This information is used later to fill in the chart below.
Month Answer

Answer: 1) Yes  2) No


3.2.6 Nutrition intervention

This group of questions should provide information on existing nutrition intervention programs and on the efficiency of health services.

37.) Question: Has the household received supplementary food through a food program during the last four weeks?

1) Yes   2) No   9) No answer
Variable code: FOODAID
This question should only be asked in regions where food assistance or feeding programs have been carried out.

38.) Question: Has one or more of the children participated in a school feeding program during the last four weeks?

1) Yes  8) Don't know
2) No   9) No answer
Variable code: SCHOOLFD
This question should only be asked if a feeding program is conducted in schools.

3.2.7 Value and norms related to gender

39.) Question: If you were to have another child, would you like it to be a boy or a girl?

1) Boy   8) Don't know
2) Girl  9) No answer
3) Boy or girl, it does not matter
Variable code: GENDPREF
In some societies girls are valued less than boys, which may indicate the presence of gender discrimination.

3.2.8 Participation in social activities

40.) Question: During the last 7 days, how often have you frequented a meeting with more than 5 persons?

Variable code: SOCIALIF
In some communities a poor nutritional status is caused by poor social relationships. This question should give information about the social life in the community. The answer is given as frequency of participation of meetings.

3.2.9 Anthropometry

The use of anthropometric data for assessing the nutritional condition of community groups has been adopted internationally as a standard practice (World Health Organization (1983), Measuring Change in Nutritional Status, WHO, Geneva). This is based on the concept that an improperly nourished body is lighter or heavier than one provided with adequate nutrients. On the one hand these data can show that weight is too high or too low in relation to height, while on the other hand they can show that the genetic potential for height has not been attained.

Anthropometric indices can be used for all members of the community, i.e., children as well as adults. Although anthropometric data on the total population can give information on both present and past nutritional, ecological, and overall development status of the community under investigation, one usually focuses on the most vulnerable segments of the population: preschool children and, to a lesser extent, women of reproductive age.

To assess the nutritional status of preschool children, ideally both weight and height (length or stature) should be measured. The term length is used for children who cannot walk yet. Their height is measured lying down. One generally refers to stature from when children can and will stand alone, i.e. around 2 years of age onward. Without precise information on age, weight by itself is of little value, as it is highly age dependent. Therefore, weight can only be interpreted when age or height is known. When surveys are being carried out among communities, in which no records of age, such as birth registration, baby welfare cards, family register, etc., are kept, it is essential that weight and height are measured. If height cannot be measured for the lack of a suitable instrument, mid-upper-arm-circumference (MUAC) should be measured.

If age is known, weight provides an excellent indicator of the child's nutritional status. With increasing age of the child, weight alone becomes progressively less sensitive in detecting undernutrition, since stunting, i.e., reduced linear growth, occurs with both chronically low food intakes and high (infectious) disease prevalence. A low age related weight (weight-for-age) may not necessarily spell danger, if the child is also short, suggesting chronic undernutrition whereby the weight may be physiologically adequate for the height of that given child, i.e., its body composition is "normal." There is evidence to suggest that a population with weight-for-age of 80% or more of the NCHS median shows only little if any association with functional disability, provided that no nutritional deficiency such as vitamin A, iodine or iron deficiency prevails.

A low weight-for-age may also be a reflection of temporary loss of fluid, such as resulting from diarrheal disease (DD), or of a combination of DD and undernutrition. When one carries out a cross-sectional survey, it is very difficult to assess how much of the low weight is due to low body tissue content caused by low energy intake and how much to loss of fluid. Considering that at any given time approximately 6% of children have diarrheal disease and that as many as 20-30%, or even more, may have suffered from a DD episode within the last seven days, the plotted weights-for-ages of a group of preschool children may in fact represent a somewhat grimmer situation, especially if the survey was carried out at the peak of the diarrheal season.

Age related height (height-for-age) alone tends to be a poor indicator of nutritional status in the very young. The older the child, the better the child reflects the combined effects of infectious disease prevalence and undernutrition on linear growth. Since, however, in an unhygienic environment, where multiple and repeated infectious are the rule, the child may have little chance to catch up on its (linear) growth deficit, height-for-age says little, if anything about the current nutritional status and whether the child is at acute risk of disease or death from malnutrition. A low height-for-age tells us only that in the past the child either has had less to eat than his/her physiological state required, i.e., suffered from chronic food shortage, or has been subjected to a high infectious disease prevalence, or both. Therefore, a low height-for-age does not necessarily indicate a disease state. It may, for instance, indicate that growth has adapted physiologically to the prevailing situation, i.e., by slowing down. A small-for-age child may be perfectly healthy and have the appropriate body composition in terms of its weight-for-height, although it appears to be malnourished both from its weight-for-age and height-for-age.

To assess nutritional status, it is therefore necessary to find out: Gender

41.) Question or observation: Is the child a boy or a girl?

1) Boy   2) Girl
Variable code: SEX Age

42.) Question or observation: When was the child born?

Variable code: BIRTHDAT Day/Month/Year
In countries in which the birth of children is registered, there is normally no difficulty in determining their age. To confirm age, in certain situations an enumerator should ask to see a baptismal or birth certificate or a clinic registration card. If the exact date of birth is not known, the month of birth can be estimated from a preestablished local "calendar of events." In agrarian societies, established festivals or agricultural events (harvesting of a particular fruit, bad weather, etc.) can be used to fix points in time. In urban areas, public holidays (e.g., national holidays) or political events (e.g., elections) can also be useful.
The age of a child has to be known or estimated in months in order to calculate the anthropometric indices.

To do this, the precise age of the child on the day of the survey must be acquired. There are often problems in this process, and consequently it is necessary to agree upon established international regulations for the calculation of age.

A child has fully ended the first month of his life on the first day of the following month (month 1). In the period of half a month before and after the completion of the first month (i.e., in the case of a 30-day month, from 16 to 45 days after birth) the child is one month old.

If for example a child were born on February 24, 1989, the child's age would be 5 months if it were visited between July 9, 1989 (9.Jul) and including August 8, 1989 (9.Aug).

As in practice errors inevitably slip into the calculation of age, it is recommended that the date of birth be asked. If the birthdate is known the Nutrition Baseline software calculates automatically the age in months. These data can also be used to check the mother's knowledge of the age of the child.

43.) Calculation: [Age of the child]

Variable code: AGE Weight

44.) Measurement: Weight of the child (00.1 kg)

Variable code: WEIGHT
The weight of children under six years of age is usually measured with a spring scale (figure 6). The prevailing first choice for this is the "Salter spring scale" (Model 235 PBW). This scale can be ordered through UNICEF (stock no.: 01-455-50, UNICEF Procurement and Assembly Centre (UNIPAC), UNICEF PLADS, Freeport, DK-2100 Copenhagen O, Denmark)

This scale can measure to the nearest 100 g the weight of children up to 25 kg. The child is weighed in a specially cut sack attached to the end of the spring scale. This model is light and easily transportable.

Figure 6. Spring scale for weighing small children (Salter spring scale)

s74.gif (624 Byte)

Another more modern scale is the UNICEF scale model 890 produced by SECA® which is a digital, solar energy operated scale with a range from 2 to 150 kg in graduations of 100 g. The scale can be ordered from UNICEF (stock no.: 01-410-00).3 The scale is powered by long-lasting lithium batteries. These will complete at least one million weighing cycles, or 400 weighings every working day for a minimum of ten years. The batteries and the electronics are in a sealed unit to withstand better humidity, heat, and dust. Additionally, the scale has a solar cell that is used only to turn the scale on and to tare the scale.

The advantages of this scale are the readability of the measurement and the ability to weigh infants and small children together with their mothers. However, the child must weigh at least 2 kg. First, the mother is weighed alone on the scale, then the automatic tare button is pressed. The child is then weighed in the mother's arms. It is important that the mother holds the child in her arms while standing in an upright position on the scale. To ensure that the mother will remain in the same position, it is recommended that the shape of two feet be drawn on the rubber mat of the scale.

The UNICEF Scale 890 is not a medically calibrated scale. It can give variations of over 100 g. Medically calibrated digital scales are much more difficult to use and more expensive and are out of the question for use in surveys. As these calibrations are set in relation to measured weights, the indicated weight can be corrected by means of calculation.

The scale weighs approximately 3-4 kg and should not be stored or transported at temperature of below 0°C or above 45°C and protected against excess humidity or wetness.
  1. In preparing for weighing, the scale should be turned on by covering the solar cell for less than one second. The display should show 188.8 first, and then 0.0.
  2. Now the mother can step on the scale and in case the weight of the mother should be noted the measurement can be taken from the display. Care should be taken that the solar cell is not covered by a foot or a long rope of the mother.
  3. For the weight measurement of the child the solar cell should be covered again for less than a second while the mother is still standing still on the scale. The display will read again 0.0. The mother can get off the scale the scale to get the child or the child can be handed over to the mother. In case the mother gets off the scale the display shows --,-. After the mother steps back onto the scale and holds the baby, only the weight of the infant will be displayed.
  4. In case another baby should be weight with the same mother (or helper), the infant be handed back and another child can be weighed without taring the scale, since the weight of the adult has been memorized automatically. If another mother (or helper) is standing on the scale, the solar cell has to be covered again for less than a second for taring.

Note: Scales can go out of adjustment during a survey. Therefore, the scale should be checked routinely, preferably at the beginning of each day. First, check the zero setting (the weight reading without any load applied) and then weigh a bucket of a known weight filled with exactly 10 L of water.

Care should be taken that during weighing the child wears no shoes and as little clothing as possible. Sometimes, however, a child must wear certain items of clothing for climatic or cultural reasons. In these cases, the weight of the clothing should be deducted from the measured weight. The average weight of clothing worn by children should be established during the pilot study. The weight is recorded to the nearest 0.1 kg.

Following the procedures outlined earlier, the infant is weighed together with his/her mother. The weight of the mother, when put into relation with her height, can be a valuable additional source of information on the social situation of women in the project area, as well as for determining the causes of malnutrition in small children. Women living in extreme poverty, underfed, frequently ill, and poorly educated, indicate a poor nutritional situation. Furthermore, underweight mothers are at high risk of giving birth to underweight babies (< 2500 g), which in turn carries a high risk of permanent poor anthropometric data for the child.

To assess the social situation of women in the project area, it is desirable to record the weight of the mother. The weight is recorded to the nearest 0.1 kg.

45.) Measurement: Weight of the mother (000.1 kg)

Variable code: MOWEIGHT

46.) Question: Is the mother pregnant?

0) No  5) Yes, 8 months
1) Yes, 0-2 months  6) Yes, 9 months
2) Yes, 3-5 months  8) Don't know
3) Yes, 6 months    9) No answer
4) Yes, 7 months

Variable code: PREGNANT Height

47.) Measurement: Height of the child (cm).

Variable code: HEIGHT

Measurement of height (length or stature) requires somewhat more effort than that of weight. Children under two years of age (i.e., up to and including 23 months) are measured in a lying position. For length measurements, a specialized wooden device ("anthropometer") should be made by a cabinet maker. An example of a device for this purpose developed by the Appropriate Health Resources and Technologies Action Group, Ltd. (AHRTAG, 85 Marylebone High Street, London, W1M 3DE England) for the WHO is illustrated in figure 7. Another option offers the "Rollametre" of the Child Growth Foundation (Raven Equipment Limited, Unit 4, Ford Farm Industrial Complex, Braintree Road, Dunmow, Essex, CM6 1HU, England)

In some studies mothers have refused to have the length of their children measured because they associate the anthropometer with a child's coffin. In these cases, success has been achieved by painting or pasting colorful flowers or animals on the wooden components.

The child is placed on its back between the slanting sides. The head should be placed so that it is against the top end. The knees should be gently pushed down by a helper. The cursor is then moved toward the child until it presses softly against the soles of the child's feet and the feet are at right angles to the legs. The length is then read in centimeters.

Figure 7. Anthropometer for the length measurement of children under two years of age.
s77.gif (4922 Byte)

If the child is over two years of age, stature is measured in a standing position. The child stands without shoes on a level floor. As shown in figure 8, it is recommended that during measuring the child leans with its back against a wall. The legs are placed against each other, as also are the heels. The buttocks, shoulder blades and head should rest against the wall. The child should look straight ahead so that an imaginary plane that would connect the eyes and ears are parallel to the floor. The arms hang loosely by the sides. The best instrument for stature measurement is the microtoise. The microtoise has to be fixed on a straight wall precisely 2.00 m above a flat floor.

The stature measurement is taken with a Stanley-Mabo "Microtoise" that can be purchased by UNICEF (stock no.: 01-144-00).3

A wooden rule or a measuring tape (preferable a tape made out of fiberglass) is placed against the wall.

Measuring tapes are sometimes calibrated in inches and centimeters on the same side. These can be confusing during measurement.
A wooden or metal right angle is employed in measuring. This instrument is placed lengthwise against the measuring tape on the wall and is pressed gently against the head so that the stature can be read on the measuring tape in cm.

Figure 8. Stature measurement for children aged two years and older.
s78.gif (1915 Byte)

The Z-scores of the following anthropometric indices are calculated automatically with the Nutrition Baseline software. The principle of these indices are explained in chapter 4.2.

48.) Calculation: [Z-score height-for-age]

Variable code: HFA
This variable shall not appear in the questionnaire

49.) Calculation: [Z-score weight-for-height]

Variable code: WFH
This variable shall not appear in the questionnaire

50.) Calculation: [Z-score weight-for-age]

Variable code: WFA
This variable shall not appear in the questionnaire

If the mother was weighed, her height should also be obtained. The same rules apply for measuring the height of a mother as for children able to stand.

51.) Measurement: Height of the mother (cm).

Variable code: MOHEIGHT

52.) Calculation: [Body mass index of mothers]

Variable code: BMI
Body weight of the mother (kg) / Body stature of the mother (m)2

This variable is calculated automatically by the Nutrition Baseline software and shall not appear in the questionnaire Mid-Upper-Arm-Circumference (MUAC)

For the measurement of the mid-upper-arm-circumference (MUAC) a special slotted "insertion" tape is used (figure 9). This tape can be purchased from UNICEF (stock no.: 01-456-00).

The best position for the enumerator measuring MUAC is sitting because the enumerator's eyes are at the level of the measurement. Before the measurement the mother should remove any clothing that covers the child's left arm. During the measurement the child can be held by the mother.

According to figure 10, the measurement of MUAC has to be taken at the middle of the left upper arm. Therefore, the enumerator must first identify the tip of the shoulder 1 ‚ and the tip of the elbow ƒ and mark the midpoint between the two tips. The child should bend its elbow to a right angle …. Once the tip of shoulder has been located the arm circumference "insertion" tape should be placed on the top of the tip at zero which is indicated by two arrows „. The tape should then be pulled down past the tip of the elbow and the number at the tip of the elbow should be read to the nearest centimeter …. The midpoint of the upper arm is determined by dividing this number by two. The midpoint should be marked on the skin of the child before removing the tape from the arm †.

Instead of the tape a string can be used for the identification of the midpoint. After the identification of the two tips one end of a string is fixed by one hand on the tip of the shoulder. With the other hand the string is pulled down to the tip of the elbow. Once the distance has been identified, the two point on the string will be brought together on the tip of the shoulder and held with the hand already holding one end of the string. The end of the doubled string is brought to the middle arm and marks the midpoint.

For the measurement of the circumference, the left arm needs to be straightened. The tape should be wrapped around the arm at the midpoint with its numbers right side up. It is important that the tape is flat around the skin ‡ and neither too tight so that it notches the skin ˆ nor too lose so that the tape looses its contact with the skin ‰.

When the tape is in the correct position on the arm with the correct tension, the measurement can be read to the nearest 0.1 cm and recorded into the questionnaire.

53.) Measurement: MUAC of the child (00.1 cm)

Variable code: MUAC

Figure 9. Tape for the measurement of the mid-upper-arm-circumference

s80.gif (1384 Byte)

Figure 10. Mid-upper-arm-circumference measurement for children

s81.gif (8179 Byte)

3.2.10 Morbidity information

A child's anthropometric data reflects the overall nutritional condition. In addition the possible presence of nutrition-related diseases with specific symptoms should be investigated (e.g., anemia, vitamin A deficiency, iodine deficiency).

If a person is identified as ill or there is reason to suspect illness, assistance should be provided. The mother, or the person responsible in case of illness, should be advised to visit the nearest health post or health assistant. Anemia

One consequence of reduced iron intake is reduced levels of red blood cells (erythrocytes) in the blood. The erythrocytes are responsible for the transportation of oxygen. The medical term for erythrocyte (Ery) deficiency is "anemia." Iron deficiency is detected by the red blood cell count or the concentration of the oxygen binding molecules of hemoglobin (Hb) in the blood.

Anemia leads to a reduced degree of physical activity in an individual and increases his/her vulnerability to infection.

To diagnose iron deficiency, it is necessary to take blood. The worldwide spread of the immune-deficiency disease AIDS has brought the risk of infection by blood transmitted diseases when taking blood to the forefront of public attention.
Before deciding to take blood samples in a baseline survey, it is necessary to weigh carefully the risk of infection. This must be done in cooperation with the local health authorities.

If in the local setting, where safety cannot be assured, it is better to skip taking blood as the risk of infection, especially with HIV or hepatitis, is too high.

If the answer to any of the following questions is "no" or "don't know," blood should not be taken!

Taking blood samples is always a critical point in a survey. It is associated with pain and crying. Most people have a natural aversion to the sight of blood. It is necessary to give a participant an understandable explanation why blood samples are needed.

Before taking the blood sample, the child should be kept on the lap of his/her mother or laid down and his sight turned away from his fingers. The blood sample should not be taken from the fingertip but from the side, as there are considerably more nerves in the fingertip and therefore more pain is felt at this location. An automatic finger prick apparatus should be used for the following reasons:

In the field the Autoclix® or the newer Softclix® apparatus has proved its reliability (Boehringer Mannheim GmbH, Germany). Packages of 200 sterile lancets can be purchased.
If despite all explanations the mother refuses to allow blood to be taken, this decision must be respected and no blood should be taken. In these cases 999 is entered into the code field.

The most exact method of measuring the hemoglobin concentration in the blood is by photometry. There are various quick test systems for this purpose.

The Compur® Minilab 1/2/3/ Photometer Systems enables the Hb concentration and Ery content in blood to be measured quickly, easily and accurately in the field (Bayer Diagnostic GmbH, Weißenseestraße 101, D-81539 München, Germany). With the Minilab 1 and 2 systems concentrations of other blood constituents, such as glucose (Gluc) and bilirubin (Bili), can also be measured. Each version of the equipment has built-in programs for the methods to obtain the results for Hb, Ery, Gluc and Bili and is calibrated by the respective Instant M End-test Cuvettes from Bayer Diagnostic + Electronic. The results are given in units and mass concentrations, which can be self-selected by the user.

The equipment is powered by either electricity (220 V - 50/60 Hz: Article No. 608925; 110 V - 50/60 Hz: Article No. 608990) or heavy duty 1.5 volt batteries. The battery power supply allows field use. The value is recorded in units of g/L.

54.) Measurement: The serum hemoglobin concentration of the child.        g/L

Variable code: HEMOGLCH in case of hemoglobin measurement of the child

HEMOGLMO in case of hemoglobin measurement of the mother

With the Minilab equipment, the Hb concentration is determined using the hemoglobin-cyanide method. To do this, the INSTANT M (40 complete tests) quick test Article No. 605749 must be used.

If the Compur rapid test is used, the Ery count can also be reported, and thus more information on the cause of anemia is obtained. Otherwise, a separate erythrocyte measurement is not recommended. Vitamin A deficiency

The most common vitamin deficiency in developing countries is that of vitamin A. This deficiency is responsible for stunted growth, greater risk of infection, various skin and eye diseases and can also lead to blindness.

Unfortunately it is currently possible to detect vitamin A deficiency at an early stage only by using expensive biochemical analysis techniques. These methods of analysis have no place in routine investigations. Consequently, it is necessary to base observations on indirect indications and signs of extreme vitamin A deficiency.

Studies have shown that the existence of a commonly used word for night blindness and the extent to which it is known in a community can indicate the presence of vitamin A deficiency. The first step is to find out whether there is a specific word for night blindness about which it is possible to ask a question.

55.) Question: Does your child stumble unusually often or have difficulties seeing or distinguishing objects at dusk or dawn?

1) Yes  8) Don't know, not sure
2) No   9) No answer
Variable code: NIGHTBLI Iodine deficiency diseases (IDD)

Another far more widespread micronutrient deficiency is iodine deficiency. It can lead to retarded mental development and cretinism.

Iodine deficiency is most easily recognized by an enlarged thyroid gland, but not every case of a enlargement can be traced to iodine deficiency. If there are many persons with swollen thyroid glands in a certain area, however, it can be concluded that there is an endemic iodine deficiency.

Since it is difficult to diagnose enlargement of a thyroid gland in small children, mothers or older schoolchildren should be checked for goiter in order to obtain information about the prevalence of possible iodine deficiency in the surveyed area.

56.) Observation: Does the mother have a visible enlarged thyroid gland or one that can be felt?

1) Yes, visibly enlarged (Grade 2)  8) Don't know, not sure
2) Yes, palpable (Grade 1)          9) No observation made
3) No (Grade 0)

The simplified classification of goiter consists of three grades:

* Mothers' thyroid should only be palpated by experienced personnel. Otherwise, thyroid gland changes should be controlled visually.

Variable code: MOGOITRE

In several countries it has been proven that IDD disappears as a public health problem if iodization of salt is carried out adequately. Therefore, the quality of the salt is a valid, simple, indirect indicator about IDD situation in the area. The next two variables are dealing with salt quality in the household.

57.) Observation: Could you please show me which kind of salt you use currently in cooking and as table salt?

1) Packaged iodized salt          3) Coarse, rock or brick salt
2) Packaged salt, without label   8) Don't know, not sure
   about iodization               9) No observation made
Variable code: SALTYPE

The presence of iodine in salt can be easily monitored using a rapid-test kit. Details about the availability of the rapid-test kids may be obtained from UNICEF6 or local authorities of Ministry of Health. The observer places one or two drops of the solution on a small salt sample (one teaspoon is adequate). The intensity of the blue color which develops indicates the salt iodine level. However, most of the test kids can detect the presence of iodine only. The expiration date of the test-kits should be checked.

58.) Measurement: Presence of iodine in tested sample

1) No color change             8) Don't know, not sure
2) Blue color change occurs    9) No observation made
Variable code: SALTIOD Diarrheal diseases

The nutrient intake of a child drops off sharply during periods of diarrhea because of loss of appetite and/or vomiting. In addition absorption of nutrients from the digestive tract can decrease as much as 70%. Therefore, there is a direct relationship between the frequency of diarrheal diseases and undernutrition, and it is important to measure the prevalence of diarrhea.

59.) Question: During the last 24 hours did your child have more than 4 liquid stools?

1) Yes   8) Don't know, not sure
2) No    9) No answer
Variable code: POINTDD
The answers to this question enable the determination of point prevalence, i.e., the proportion of children suffering diarrheal disease at a certain point in time.

60.) Question: Has the child suffered from diarrheal disease during the last 7 days?

1) Yes   8) Don't know, not sure
2) No    9) No answer
Variable code: PERIODDD
The answers to this question will enable the determination of the period prevalence of diarrheal disease. The period prevalence thus derived is not identical to the above observed point prevalence, which is the reason that specific questions should be asked concerning both indicators. Many studies have shown that a mother's recollection drops sharply after a period of one week. Therefore, the question is not posed concerning a longer period. Acute respiratory infections (ARI)

Several studies have demonstrated a link between undernutrition and acute respiratory infections (ARI). Due to the widespread presence of these infectious diseases it is necessary to determine their prevalence.

61.) Observation: Is the child currently suffering from running nose, cough, cold, earache, or sore throat?

1) Yes  8) Don't know, not sure
2) No   9) No answer
Variable code: POINTARI

The survey team should determine the response directly from the child and not by asking the mother. If the child suffers at least one of the following symptoms: common cold, earache, sore throat, cough, raspy and/or rapid breathing (breathing rate > 50 breaths per minute), acute respiratory infection is likely.

It is not necessary to go into any further classification of ARI. In case of any sign of disease, the child should be send to the next health service for an adequate treatment and no therapy can and should be provided by a non-medical enumerator.

The answers to this question enable the determination of the point prevalence of acute respiratory infections.

62.) Question: Has the child suffered from running nose, cough, cold, sore throat or earache during the last 7 days?

1) Yes   8) Don't know, not sure
2) No    9) No answer
Variable code: PERIOARI

The answers to this question will enable the determination of the period prevalence of acute respiratory infections. The answer should be given by the mother. Other infectious diseases

As explained earlier, infectious diseases can be responsible for poor nutritional status. In developing countries, the most prominent diseases in this area are acute respiratory infections and diarrheal diseases. In some regions measles or other infectious diseases such as malaria, tuberculosis, etc. are endemic and thus negatively influence the nutritional status.
In some areas special attention is needed concerning AIDS in small children. As the disease progresses the child's body constantly wastes away, resulting in visibly worsening anthropometric data.

Problems with infections in the surveyed region must be identified during the pilot study and the questions on the questionnaire adapted to the respective problems.

The following question applicable to measles serves as an example of a supplementary question for the measurement of period prevalence of infectious diseases.

63.) Question: Has the child suffered (e.g., measles) within the past year?

1) Yes  8) Don't know, not sure
2) No   9) No answer
Variable code: PERIODIS Mortality of preschool children

The number of children that were born alive and died is a very important indicator of the general living and health conditions of the community. Although this number is not identical with the statistics of infant or child mortality, it gives useful indirect information about child mortality.

64.) Question: How many children have you born alive?

Of these ..... children how many are still alive today?
How many of the .... children that died, died before the age of 5 years?

Variable code: CHILDDTH

It is advisable not to ask straight forward how many children died before the age of 5 years, but for the sake of clarity it is better to lead with several questions to the wanted information.

3.2.11 Infant nutrition

A common cause of nutritional deficiencies in small children can be found in inadequate infant nutrition. Improper feeding practices during the first two years of life can lead to many functional disorders and irreversible stunting.

Because of the importance of infant nutrition to nutritional status, a nutritional baseline survey should contain a number of questions on this area of nutritional practices. Breast-feeding practices

Breast-feeding is still the prevailing practice in many communities. It is known that there is no more suitable source of nutrients for an infant up to four months of age than exclusive breast-feeding. Early weaning of babies drastically increases the risk of contracting infectious diseases.

65.) Question: Have you breast-fed your child during the last 24 hours?

1) Yes   2) No   9) No answer
Variable code: BREASTFE

66.) Question: If your child is not currently breast-fed, how long did you breast-feed your child?

Variable code: BREASTDU

If currently breast-fed the response is 66; if never breast-fed the response is 00. Otherwise, the response is reported in months.

67.) Question: Did your child receive colostrum?

1) Yes  8) Don't know
2) No   9) No answer
Variable code: COLOSTRU

In all cultures, a local word is used for colostrum. This term should be identified to be used for questioning. Colostrum is the first breast milk of the mother that is fed to the child. Due to its high content of nutrients and immunoglobulins the first milk has a more intense color than the later breast milk. As a result, in many cultures this milk is often diluted with other fluids or even discarded.

68.) Question: How many hours after birth did you put your baby to the breast?

1) Immediately  4) After 12 hours
2) 1-4 hours    8) Don't know
3) 5-12 hours   9) No answer
Variable code: STARTBF

69.) Question: Did your child receive any other fluid besides breast milk during the first days after birth?

1) Yes  8) Don't know
2) No   9) No answer
Variable code: ADDITBF

In some cases, the newborn infant receives other fluids beside breast milk. Thus after the first hours after birth, the infant has not yet produced larger amounts of protective enzymes and acids in the intestine and the membranes are still vulnerable to be penetrated by larger molecules. Therefore, the administration of other fluids than colostrum during the first hours of life carry a high risk of infectious diseases.

70.) Question: Within the last 24 hours has your child drunk anything in addition to breast milk?

1) Yes  8) Don't know
2) No   9) No answer
Variable code: ADDDRINK

71.) Question: If your child receives solid food, at what age did you begin feeding the child solid food?

1) < 4 months 8) Don't know
2) > 6 months 9) No answer
3) 4 - 6 months
Variable code: AGESOLID Supplementary feeding and weaning practices

Before the survey it must be established beyond doubt that mothers are prepared to answer the following questions. For example, in a survey in Pakistan questions concerning types of feeding for small children were considered an affront, and the interviewee subsequently broke off the dialogue.

72.) Question: Regardless of whether your child is breast-fed or not, how often has your child eaten any solid food in the last 24 hours?

1) Yes  8) Don't know
2) No   9) No answer
Variable code: EATSOLID

The response is recorded in number of occurrences.

73.) Question: In the last 24 hours has your child eaten fruits or vegetables?

1) Yes   2) No   9) No answer
Variable code: EATGREEN

74.) Question: During the past 24 hours has your child been bottle-fed?

1) Yes   2) No   9) No answer
Variable code: BOTTLE

75.) Question: Regardless of whether your child is breast-fed or not, how often was your child given something to eat yesterday?

1) Yes  8) Don't know
2) No   9) No answer
Variable code: EATFREQU

3.2.12 Formal under-fives health services

The following questions should give some information about the accessibility and quality of formal under-fives health services related to nutrition. Weight monitoring

Surveying the proportion of households possessing weight monitoring cards, it should be possible to determine the effectiveness of under-fives health services in the area of nutritional intervention.

76.) Observation: Does the mother possess a weight chart for the child?

1) Yes   2) No    9) No observation possible
Variable code: WHTCHART
The mother should present the child's weight chart.

77.) Observation: Has the child been weighed during the first year of life at intervals of not more than 3 months?

1) Yes   2) No   9) No observation possible
Variable code: WEIGHING Immunization

To arrive at a second indicator for the effectiveness of under-fives health services, a survey should establish the proportion of households possessing immunization cards, and also the proportion of children immunized after the age given in the immunization plan.

78.) Observation: Does the mother possess an immunization record for the child?

1) Yes   2) No
Variable code: IMMUNCRD

3.2.13 Acceptance of a survey

Finally, an enumerator should ask the community members about their opinions concerning the type and format of the survey.

In cultures in which community members are comfortable with making critical statements, the survey may be used to obtain information on the opinion of the interviewees about the implementation of the survey. Otherwise, community opinion should be ascertained from unstructured interviews.

Valuable information can be obtained for the planning of future interventions from the evaluation of this survey. The degree of participation in the survey can be a useful indicator of the acceptance of a survey. If only a few people want to be informed about the results of the survey, then only a few will be interested in participating in the planning and implementation of the intervention. Furthermore, asking about the acceptance of the survey will instill confidence that the institution in charge of the nutrition survey is genuinely interested in the participation of the target community.

79.) Question: Would you be prepared to take part again in a similar survey?

1) Yes
2) Only if ____________________________________________
3) No, because ________________________________________
Variable code: SURVPART

80.) Question: Are you interested in a discussion on the findings?

1) Yes 3) No
2) Eventually 8) Don't know
Variable code: SURVDISC

If the community is inadequately organized and/or the members attend different activities, which often happens in newly developed urban areas, it is advisable to ask during an interview about the most convenient time and place for a future meeting with members of the community.

81.) Question: Which would be for you the most opportune day of the week to receive the information about the survey and to discuss it with your neighbors?

1) Monday     5) Friday
2) Tuesday    6) Saturday
3) Wednesday  7) Sunday
4) Thursday   8) Don't know
Variable code: INFODAY

These questions on the day of the week and the time of day are only provided here as examples. It is definitely not applicable in all cultural situations. More often, these questions must be asked in relation to normal living habits.

82.) Question: What would be the best time for the discussion with your neighbors?

1 - 24 hours 88) Don't know 99) No answer
Variable code: INFOTIME

3.3 Reliability check for a survey

The reliability of data collection in a survey essentially depends on

The preceding chapters deal with these points in detail. Nevertheless, it is only possible to assess objectively the reliability of collected data if the recorded measurements, observations and answers are compared with the actual situation and the proportion of correct to incorrect values obtained. However, it is often difficult to ascertain the "real" situation.

In practice, a supervisor is appointed to ascertain the "real" situation in a survey. The rationale for this is that a supervisor is better trained and has more experience than the enumerators. The degree of variance in the collected data provides an indication on the reliability of data collection (for more details, see chapter 4.6.).

During the course of a survey, the supervisor collects data on the variables with the highest risk of yielding unreliable results. The selection of these variables should be made after the pilot study (see sub-chapter 3.1.6). In all cases, the supervisor should take measurements of height and weight so that anthropometric data can be calculated and compared to that obtained by the survey team. An example of a questionnaire for a supervisor is presented as an appendix in sub-chapter 6.1.4.

Altogether about 10% of the households visited by enumerators should be selected at random for a cross-check. In addition to serving the purpose of data collection for a reliability check, the cross check also provides a supervisor an excellent opportunity to observe first hand the difficulties of the enumerators in recording data in order to be able to introduce subsequent countermeasures during the survey.