Previous PageTable Of Contents

Part III. Methods of social analysis

Part III. Methods of social analysis

Chapter 5 Collecting and using social data

This chapter is concerned with the collection and use of data for social planning purposes. The first two sections look at why data is needed and how to identify the kind of data needed for a particular planning purpose. The last two sections then examine methods of collecting and using the two main types of data: primary and secondary.

5.1 The need for data

Data is a basic requirement for any form of planning, since planning involves making calculated decisions and such decisions must be based on adequate and appropriate information, or 'data'. Data is needed at all stages of the planning process, including the initial identification of a problem or objective, the identification and appraisal of alternative policies, programmes or projects, monitoring the implementation process, and finally the evaluation of the impact of the plan.

Social planning is no exception in this respect. However, both the type of data needed and the problems involved in collecting and using it vary from one type of planning to another. This chapter is therefore concerned with what might be called 'social' data -in other words, the sort of data needed specifically for the social aspects of planning -and in particular with the kind of social data needed for rural area planning.

5.2 Identifying data needs

One of the most fundamental but difficult tasks in any form of planning is to decide what data to collect. If this is not done before data collection starts, a great deal of time is wasted collecting data which is never used, and some essential data is inevitably ignored. This section looks briefly at a number of questions regarding the quantity, quality, type, form and source of data which it is necessary to ask before starting to collect data.

How much data?

The first step in determining how much data to collect is to consider exactly what data one needs in order to achieve the particular objective of the planning exercise. In other words, one should not just go out and collect all the data which may be vaguely relevant. However, it is never possible to obtain as much data as one would ideally like, especially at district level, where the resources for collecting and analyzing data are generally severely limited. It is therefore necessary to reconcile the amount of data one would like to have with what it is actually practical to obtain. The aim should be to collect the minimum amount of information needed to achieve the particular objective. And if there is any serious gap in information, one should be aware of any limitations this may have in terms of the accuracy or comprehensiveness of the decisions made.

This is particularly important in the case of social data, some of which is very difficult to collect quickly and cheaply. For example, one must have information on the social structure of a village in order to plan an agricultural project for the village. The implications of not doing so were demonstrated by the example of the vegetable garden project in Gondwanaland, described in Box 3.3. However, this does not mean that one needs to know everything there is about the village. Moreover, there are several different ways of obtaining the information required, some of which require more resources than others. This point will be discussed later.

What subject matter should be included?

Social data for rural area planning includes data on any aspect of the area defined as 'social' in Chapter 1. It thus includes information on the various 'social characteristics' of the area (eg. demography, ethno-linguistic characteristics, social structure, inheritance systems, religious and cultural beliefs and practices, and individual or group attitudes), the general quality of life, the quantity and quality of social services, and social justice.

However, in order to determine what kind of data should be obtained for any particular planning exercise, it is again necessary to consider the objective of the exercise. In some cases there is a need for an overview of all social aspects, as for example when one wishes to assess the general level of social development in the area. Box 3.2, which provides a list of 'social indicators' for Gondwanaland District, illustrates the kind of subject matter which might be included in such a case. Since this example will be used to illustrate various points in this chapter, it has been reproduced here as Box 5.1. However, in many cases there is a more specific need for data on selected social characteristics or issues relevant to the particular planning exercise.

Quantitative or qualitative data?

Quantitative data is data which can be expressed in numerical form, while qualitative data is expressed in the form of verbal descriptions rather than numbers. In Box 5.1, for example, all the data on quality of life and social services is quantitative while that on social justice is a mixture of quantitative and qualitative data.

BOX 5.1

GONDWANALAND DISTRICT: SOME SOCIAL INDICATORS

 

Quality of life

 

Infant mortality (per 1000 live births) 1990:

68

Average proportion of children under 5 years attending clinics who were malnourised in 1992:

9%

Adult literacy 1990: male: 62%; female:

53%

Average land cultivated per household 1991:

1.9 hectares

Average number of livestock units owned per household 1991:

9

Average annual agricultural income per household 1991:

NK$ 560

Proportion of households owning 1990:

 

motor vehicle: 9%

plough: 57%

bicycle: 28%

radio: 25%

 

Social services

  
   

Proportion of primary school age children at school 1992:

 

male: 89%

female: 78%

Proportion of 1991 primary school leavers going on to secondary school:

  

male: 48%;

female: 35%

Proportion of population within 5 kms. of primary school 1992:

85%

Proportion of population within 10 kms. of secondary school 1992:

78%

Average class size 1992: primary: 38; secondary:

34

Proportion of population within 10 kms. of clinic 1992:

82%

Proportion of population served by mobile MCH clinic 1992:

73%

Population per doctor 1992:

approx. 111,000

Population per hospital bed 1992:

1485

Proportion of households with water supply 1990:

 

inside house: 6%

less than 100 metres: 24%

101-1000 metres: 38%

over 1000 metres: 32%

 

Proportion of households with toilets 1990:

flush: 5%;

pit: 57%;

none: 38%

Social justice

Inequality between areas: Some of the indicators of quality of life and access to social services listed above are available only for the district as a whole. However, where a breakdown does exist, it suggests significant variations in some indicators. The most noticeable observation is that the standard of living in Zone V is far lower than that elsewhere (see Box 3.4 for details).

Inequality between households within areas: There is very little available statistical information. However, data from the 1991 National Sample Agricultural Survey suggests the following variations

(a) average area cultivated per household:

hectares

nil

0.1-1.0

1.1-2.0

2.1-3.0

3.1-5.0

5+

% h'holds

5

29

33

19

9

5

(b) average number of livestock units per household:

livestock units

nil

1-5 6-10

11-15

16-20

20+

 

% h'holds

12

24

26

19

13

6

(c) average annual agricultural income per household:

income $

0-100

101-250

251-500

501-750

751-1000

1000+

% h'holds

8

16

32

24

12

8

Informal observations by agricultural extension workers suggest that these I figures are probably fairly representative of the district as a whole, excluding Zone V.

Gender relations: There is a traditional division of labour between men and women in relation to agricultural and other activities. This tends to discriminate against women (and girls), in that they have fewer opportunities to engage in income earning activities, do most of the household work are less likely to be involved in decision-making (especially outside the family) and (as the above statistics indicate) are likely to be less educated. However, practices vary considerably from one part of the district to another and from one household to another.

Participation in 'development': 76% of the eligible population voted in the 1990 general election and 53% in the 1991 District Council elections. The Council tends to be dominated by a small number of councillors and is divided along Party lines. There is only one woman councillor. Each village is required to have a village development committee (VDC) but in many villages the VDC conflicts with traditional leadership structures and in some it is lime more than a vehicle for party politics. The Kurda people in Zone V have little or no influence on decision-making in the district (see Box 3.4 for details).

Sources of data

Reports of 1990 National Population Census

Report of 1991 National Sample Agricultural Survey

District records of Ministries of Education and Health

Reports of agricultural extension staff

General knowledge/observation

It is sometimes assumed that social data is more likely to be qualitative in form than other kinds of data, particularly economic data. There is some truth in this assumption, in that social data is more likely to include information which is best expressed in qualitative rather than quantitative form. However, in most cases, the best way to present a clear picture of an issue or situation, whether it be economic or social, is through a combination of quantitative and qualitative data.

However, this does not mean that the two kinds of data are interchangeable, since each has its advantages and disadvantages. When choosing whether to collect quantitative or qualitative data, there are four main factors which need to be considered:

How objective should the data be?

A distinction is often made between objective data, which is independent of the attitudes or prejudices of the people involved in collecting or providing the information (eg. the interviewers or interviewees), and subjective data, which is influenced by such attitudes or prejudices. In practice, however, data is very seldom totally objective, because those involved almost inevitably influence the data collection exercise in some way or other. In fact, the initial process of deciding what data to collect and what not to collect immediately introduces an element of bias into the exercise. For example, the social indicators in Box 5.1 give a picture of social development in Gondwanaland which is biased in favour of, firstly, those aspects of social development which the planners (as opposed, say, to the general population) think are most important and, secondly, those aspects for which data is available. In reality, therefore, it is not a question of whether the data is objective or subjective but of the degree of objectivity.

There is a tendency to assume that qualitative data is less objective than quantitative data. However, although this is often the case, it is not necessarily true. There is plenty of room for subjectivity in the collection of quantitative data, especially if some sort of survey is involved, since the data will be influenced by the prejudices and errors of the people collecting and providing the information. For example, since much of the quantitative data in Box 5.1 was obtained from the reports of the 1990 Population Census and the 1991 Sample Agricultural Survey, its objectivity is dependent on the honesty with which interviewees answered questions, the diligence and integrity of the interviewers and, in the case of the Sample Survey, the way in which households were selected. In other words, quantitative data may look more objective, because it is expressed in numerical form, but in fact its objectivity depends on the way in which it is collected and analyzed.

Similarly, there is a tendency to assume that objective data is 'better' than subjective data, because it is not 'distorted' by human bias or prejudice. Once again this is an assumption which is often but not necessarily true. A great deal depends on the purpose for which the data is required and how it is used. In some cases, particularly when the main purpose is to gain an understanding of people's attitudes and behaviour, one is actually looking for data which reflects the biases and prejudices of the interviewees. Moreover, even when this is not the case, it is generally better to use qualitative data which is known to be subjective than quantitative data which appears to be objective but in actual fact is not. In such cases, the important point is to be aware of the limitations of the data one is using and, therefore, any decisions one may make on the basis thereof. The implications of this in terms of the accuracy of data will be considered later.

How should the data be disaggregated?

Data can be broken down, or disaggregated, in various ways, and before starting to collect data it is necessary to consider how it should be disaggregated, since this will affect the form in which it is collected. The two main ways in which social data for rural area planning are likely to be disaggregated are on the basis of geographical area and social group.

Disaggregation of data on the basis of geographical area is used to distinguish and compare the characteristics of particular geographical areas. For example, most of the data in Box 5.1 relates to Gondwanaland District as a whole, and is thus useful for indicating the level of social development in the district as a whole and comparing it with other districts. However, for most district planning purposes it is necessary to distinguish between different parts of the district, either to prioritize areas on the basis of need or potential or to focus on a particular part of the district where, for example, a particular project is being planned. A preliminary indication of the degree of geographical variation within Gondwanaland is given in the section on 'inequality between areas' in Box 5.1, while the information on Muriwana and Wiriwana villages in Box 3.1 (Chapter 3) is an example of data related to specific parts of a district - in this case, two villages.

Disaggregation of data on the basis of social groups is used to identify variations or inequalities between social groups or to focus attention on any particular group. This is particularly important in social planning, since one of its objectives is to reduce inequalities between social groups. The sections on 'inequality between households' and 'gender relations' in Box 5.1 give some indication of the type of data needed for such purposes - and the difficulty of obtaining such data. This will be discussed further in Chapter 8, which focuses specifically on planning for disadvantaged social groups.

When should data be collected?

Data may be collected at various times, depending on the type of data and the purpose for which it is required. There are three main possibilities: a 'one-off' approach, a 'time series' approach, and a 'before and after' approach.

The 'one-off, approach is one where data is collected and presented for one particular point in time only, as in the case of most of the indicators in Box 5.1. This is sufficient if one only needs to know about the present situation or if, as in Box 5.1, one wants a 'snapshot' of the situation, possibly to make a comparison with other situations. For example, the data in Box 5.1 could be used to compare the level of social development in Gondwanaland with that in other districts at the same point in time.

The 'time series' approach involves collecting data at regular intervals over a predetermined period of time. Much of the data in Box 5.1 is, in fact, extracted from time series data, although the intervals vary considerably. For example, population censuses are usually conducted every ten years, while health and education statistics are usually compiled on an annual basis. Time series data is necessary if one needs to know about historical trends or if there is likely to be considerable variation over time and so a 'one-off' survey could be misleading.

The 'before and after' approach involves two major data collection exercises: one before an anticipated change or event and one afterwards. This approach is usually used to evaluate the impact of a particular policy, programme or project. It is obviously important to plan such an evaluation before implementation starts, so that one can undertake the 'before' study. In many cases, planners do not think about evaluation until implementation is complete, by which time one can only do an 'after' study. This kind of data collection will be considered further in Chapter 7, which focuses on project appraisal and evaluation.

How accurate should data be?

The question of accuracy has already been touched upon on a number of occasions, since one of the main concerns when addressing all the earlier questions is to ensure that the information obtained is as accurate as possible At this point, therefore, it is merely necessary to make a few general comments about the degree of accuracy necessary. The situation is similar to that of the quantity of data, in that the information one is able to obtain is seldom as accurate as one would wish. The aim, therefore, should be to achieve the minimum degree of accuracy necessary for the particular purpose concerned and to be aware of any possible inaccuracies in the data which may affect the validity of the conclusions reached or decisions made.

There is a tendency, particularly among economists, to assume that social data is less accurate than other kinds of data, including economic data, because it is more likely to be qualitative and/or subjective in nature. However, there are two major flaws in this argument. Firstly, as already indicated, social data can and should include both quantitative and qualitative data and data with varying degrees of objectivity, depending on the type of data and the purpose for which it is required. Secondly, data which is qualitative and subjective is not necessarily less accurate, since accuracy depends not on the type of data but on the way in which it is collected and analyzed. As already indicated, there is plenty of room for error in the collection and analysis of quantitative data. Furthermore, since such errors are difficult to identify, especially when the data is presented in its final form, there is a very real risk that conclusions will be reached or decisions made on the assumption that the data is accurate when in fact it is not.

Primary or secondary data?

Data can be obtained from many different sources. However, it is conventional - and useful for the present purpose - to divide these into two main categories: primary and secondary. Primary data is data which is collected for the particular planning purpose in question, while secondary data is that which has already been collected (either for a previous planning exercise or as part of a general data collection programme) and is merely utilized for the present planning purpose.

In any form of planning, it is advisable to assess the availability of secondary data before embarking upon a primary data collection exercise, since the latter is expensive in terms of time, money and manpower. This is particularly true at district level, where such resources tend to be particularly scarce. However, if the necessary secondary data is not available, or not available in an appropriate form, or if it would take longer to explore the secondary data possibilities than to go out and collect the data oneself, it is necessary to plan a primary data collection programme. In practice, most planning activities tend to utilize a combination of primary and secondary data.

The last two sections of this chapter examine in some depth the collection and use of these two types of data, beginning with secondary data, since (as already indicated) it is advisable to utilize any available secondary data before starting to collect primary data.

5.3 The collection and use of secondary data

Secondary data may take many forms and its quality and value for rural area planning varies greatly. In some countries, districts are encouraged to maintain a 'store' of basic secondary data about the district which can be used as and when needed for planning purposes. In order to be effective, such a store should be kept in a central place, such as a planning office or conference room, and it should be up-dated regularly. In many countries, no such store exists and it is therefore necessary to collect the secondary data individually for each planning exercise. However, in such cases, it is advisable for those involved in planning on a regular basis to familiarize themselves with the kinds of secondary data available so that they know where to go when the need arises.

This section of the chapter looks at five different types of secondary data, all of which are likely to be needed for the social aspects of rural area planning and are relatively easily available at district level. They are: demographic data; sample agricultural household surveys; official government records; maps; and general knowledge. These are not the only types of secondary data which may be used, but they give some indication of the range of information available from secondary sources - and the problems likely to be encountered in collecting and using such data.

Demographic data

Demographic data is data on the size and structure of a population, including the total population, household size, distribution by age and sex, past and future rates of growth, fertility and mortality, migration and population density. This sort of information is frequently required for a variety of different planning purposes at district level. Sometimes the need is for aggregate data for the district as whole, sometimes for data disaggregated on the basis of administrative subdivision or social group, and sometimes for data just for one area or community.

It is difficult to get accurate demographic information without a detailed house-to-house survey, which is expensive in terms of time, money and manpower. Therefore, unless the information is needed only for a small area where it is feasible to undertake such a survey, it is normally necessary to rely on secondary data, although some 'short-cut' methods of collecting reasonably accurate data on a larger scale will be described in section 5.4.

The main secondary source of demographic data is national population censuses. These are normally undertaken (if political and economic circumstances permit) approximately every ten years, and in some countries the information is supplemented by one or more inter-censal sample surveys. In most cases, such censuses also include some other basic household data, such as occupations, education and access to services such as water and sanitation. In Box 5.1, for example, the data on infant mortality, household assets and access to water and sanitation was derived from national census data.

There are two main problems in using census data at district level. Firstly, the data is not always disaggregated on the same basis as that needed for district planning purposes. For example, census enumeration areas do not always correspond to local administrative units (such as villages or wards) and, although the data is broken down by sex and age, it is seldom available on the basis of any other social criteria, such as ethnic group or social class. And secondly, it is often not sufficiently up-to-date. It is usually several years before the full results of a population census are available to the public, by which time they are already somewhat out-of-date, and then there is a ten-year gap before the next census data becomes available. This presents particular problems in districts where the population is growing rapidly or unevenly, since it is difficult to make accurate projections on the basis of the information available.

It is not possible to look in detail at all aspects of demographic data here. However, Boxes 5.2 and 5.3 illustrate two possible applications of census data for planning purposes in Gondwanaland, both of which are likely to be fairly widely used in any district.

Box 5.2 presents the data on population density by ward used to compile Map 3 in Chapter 2. It includes the following information: the area and 1990 population of each ward, as recorded in the 1990 National Population Census; the estimated present (1993) population of each ward, which is calculated by multiplying by the average annual growth rate for the district as a whole over the previous ten-year period (which is available from the Census); and the population density, which is the 1993 population divided by the area. In this case, the exercise was relatively easy because the census data was disaggregated on the basis of wards, which are the main units of administration and planning in the district. Moreover, the population projections are inevitably approximate, because they do not make any allowance for changes in the average rate of population growth or for variations in the rate of growth from one part of the district to another.

BOX 5.2

GONDWANALAND DISTRICT: POPULATION DENSITY BY WARD

WARD

AREA (SQ.KM.)

POP'N 1990(1)

EST.POP'N 1993 (2)

POP'N PER SQ.KM. 1993

1

303

9,927

10,847

35.8

2

208

7,632

8,340

40.1

3

478

9,317

10,181

21.3

4

492

8,780

9,594

19.5

5

545

10,424

11,390

20.9

6

432

9,330

10,195

23.6

7

221

10,132

11,072

50.1

8

199

9,525

10,408

52.3

9

296

11,106

12,136

41.0

10

299

11,437

12,498

41.8

11

320

11,743

12,832

40.1

12

350

10,057

10,990

31.4

13

151

7,725

8,441

55.9

14

255

11,457

12,520

49.1

15

285

11,015

12,037

42.2

16

180

8,747

9,558

53.1

17

236

7,386

8,071

34.2

18

230

10,166

11,109

48.3

19

303

9,695

10,594

35.0

20

158

9,008

9,843

62.3

21

130

8,149

8,905

68.5

22

273

8,994

9,828

36.0

23

225

11,531

12,600

56.0

24

278

11,576

12,649

45.5

25

281

11,855

12,954

46.1

26

196

10,241

11,191

57.1

27

380

11,893

12,996

34.2

28

220

7,731

8,448

38.4

29

287

9,797

10,705

37.3

30

199

9,087

9,930

49.9

31

206

9,633

10,526

51.1

32

182

8,694

9,500

52.2

33

425

9,840

10,752

25.3

34

280

6,201

6,776

24.2

35

296

7,693

8,406

28.4

36

415

9,608

10,499

25.3

37

331

8,331

9,103

27.5

38

201

5,500

6,010

29.9

39

256

10,682

11,673

45.6

40

460

5,851

6,394

13.9

41

586

$098

8,849

15.1

42

1,003

6,885

7,523

7.5

43

1,116

7,966

8,705

7.8

44

965

7,242

7,913

8.2

Total

15,132 (3)

407,687

445,491

29.4 (3)

(1) 1990 National Population Census

(2) 1990 Census figure + 3% per annum increase for 3 years

(3) Excludes forest reserve

Box 5.3 shows the population of Gondwanaland District as a whole subdivided by age and sex, based on the 1990 Census data. The data is presented in two forms: a table and a diagram known as a 'population pyramid'. It then shows how this data can be used to calculate the numbers of boys and girls of primary school age in 1993. This in turn can be used, together with educational statistics, to calculate the proportion of boys and girls in this age group who actually attend school, which is (as indicated in Box 5.1) an important social indicator. Similar calculations can be made to determine the number of males and/or females in other age categories (eg. secondary school age, economically active, dependent elderly).

Sample agricultural household surveys

In many countries sample agricultural household surveys are undertaken at regular intervals, by either the ministry responsible for agriculture or a national statistical agency, to produce basic data on the volume and methods of agricultural production. There are several different methods used for such surveys. However, in most cases households are selected from different agro-economic zones and information is collected over a full agricultural year, in order to obtain information on production methods. The information collected usually includes household size, area cultivated by different crops, inputs (including labour), yields, crop sales and income, and (although sometimes as a separate livestock survey) the number of livestock owned by type and the number sold.

BOX 5.3

GONDWANALAND DISTRICT: POPULATION BY AGE AND SEX 1990

AGE GROUP

MALES

FEMALES

TOTAL

(YEARS)

NO.

%

NO.

%

NO.

%

0-4

35,877

8.8

35,841

8.8

71,718

17.6

5-9

29,354

7.2

30,964

7.6

60,318

148

10-14

20,384

5.0

23,646

5.8

44,030

10.8

15-19

20,792

5.1

19,161

4.7

39,953

9.8

20-24

18,350

4.5

19,569

4.8

37,919

9.3

25-29

18,342

4.5

18,346

45

36,688

9.0

30-34

13,453

3.3

14,269

3.5

27,722

6.8

35-39

11,415

28

12,638

3.1

24,053

5.9

40-44

8,154

2.0

10,600

2.6

18,754

4.6

45-49

6,523

1.6

8,561

2.1

15,084

3.7

50-54

4,485

1.1

5,708

1.4

10,193

2.5

55-59

2,854

0.7

4,468

1.1

7,322

1.8

60-64

2,446

0.6

3,262

0.8

5,708

1.4

65-69

1,631

0.4

1,691

0.4

3,322

0.8

70-74

1,226

0.3

1,235

0.3

2,461

0.6

75+

1,219

0.3

1,223

0.3

2,442

0.6

TOTAL

196,505

48.2

211,182

51.8

407,687

100.0

Calculation of primary school age population 1993

Primary school age = 6-12 years inclusive

Children aged 6-12 years in 1993 were 3-9 years in 1990

Thus, potential primary school age population = all those aged 5-9

years and approx. 2/5 of those aged 04 years in 1990=

males:

29,354 + 14,351 = 43,705 (approx.)

females:

30,964 + 14,336 = 45,300 (approx.)

total:

60,318 + 28,687 = 89,005 (approx.)

But some of these children will have died between 1990 and 1993.

After adjusting for mortality rate of 9 per 1000 each year, estimated primary school age population =

males:

42,536 (= approx. 42,500)

females:

44,088 (= approx. 44,000)

total:

86,624 (= approx. 86,500).

Population by age and sex - 1990

This sort of information is much needed for district planning purposes, particularly in the agricultural sector, and - like demographic data - it is the sort of information which it is difficult to obtain accurately without a detailed household survey. However, the extent to which such surveys can be used at district level varies, depending on the size of the sample, the way in which it is selected, and the way in which the results are disaggregated. In most cases, this kind of survey is undertaken for national rather than district planning purposes. Consequently, the sample is relatively small and nationwide agro-economic zones usually provide the basis for both selecting the sample and disaggregating the data. However, in some countries where the importance of district level agricultural planning is recognized, such surveys are designed in such a way that they provide statistically valid data at least for the district as a whole and sometimes for particular agroeconomic or administrative areas within districts.

In Box 5.1, the information on land cultivated, livestock units and agricultural income presented in the sections on 'quality of life' end 'inequality between households within areas' were derived from a National Sample Agricultural Survey. In this case, the survey provided data which is statistically valid for the district as a whole but not for individual subdivisions. Consequently, it is useful for giving a general picture of agriculture in the district as a whole, which can in turn be used to make comparisons between districts or (if such surveys are undertaken regularly) over time. However, it is of limited value for detailed agricultural planning. For such purposes, it will be necessary either to rely on reports or casual observation by agricultural staff (see below) or to undertake special surveys (see section 5.4).

Official government records

One of the most useful potential sources of information for district planning purposes is the official records which all government agencies at district level are required to keep and submit periodically to their national headquarters. These records usually take two main forms: reports on activities and problems and statistical returns. This information is primarily intended for national planning and monitoring purposes and district staff tend to regard the maintenance of such records as an unpleasant and largely unnecessary chore. However, a considerable part of the information can usually be fruitfully used at district level. Moreover, with a little ingenuity, district staff can modify the official recording procedures to collect any additional information they may themselves need while still meeting the requirements of their head offices.

All the data on education and health services in Box 5.1 is derived from statistical returns by district representatives of the Ministries of Education and Health. In some cases (eg. the proportion of primary school leavers going on to secondary school and the average class size), such information can be obtained directly from these returns. In other cases, some form of analysis is normally required. For example, in order to know the proportion of primary school age children attending school it is necessary to calculate the number of primary school age children, as demonstrated in Box 5.3. Similarly, in order to know the proportion of the population within a specific radius of a particular education or health facility, it is necessary to locate these facilities on a map and relate this to information on population distribution. This is a somewhat more complex exercise, which will be described in Chapter 6, in the section on education planning.

Maps

Maps are a very useful way of presenting data for district planning, especially as a means of conveying information or stimulating discussion at meetings of civil servants, local politicians and/or the general public. Since they portray data in a visual form, they are simpler and clearer than written reports or statistical tables and even people who are not familiar with maps can quickly learn to understand them. They are particularly useful in indicating the degree and form of variation within the district and the relationship between different sectoral activities within a particular geographical area.

Secondary data is often already available in map form, although sometimes the scale of the map is inappropriate for district planning purposes and so the data has to be transcribed onto another map. For example, published maps of physical characteristics (relief, rainfall, geology, soils, etc.) and of the main settlements and communications (roads, railways) are usually available on a nationwide basis, while the district offices of government agencies frequently have their own, hand-drawn maps showing the particular infrastructure or services with which they are concerned (eg. schools, health facilities, water supplies, roads). And data which is not already in map form, can often be presented in such a way without a great deal of effort.

The maps of Gondwanaland in Chapter 2 indicate some of the data which can be presented in the form of maps. Most of the data on these maps is likely already to be available in map form. However, the map of population distribution (Map 3) might have to be compiled by the planner from data such as that in Box 5.2. Further examples of the use of maps will be given in Chapter 6.

General knowledge

Another very valuable source of information for planning at district level is the considerable amount of general knowledge about the district which those involved in planning inevitably have, especially if they have lived or worked in the area for many years. This sort of information is frequently used for planning purposes, but often more or less unconsciously. For example, when a government officer makes an initial proposal for a project which he (or she) considers necessary or contributes to a discussion on someone else's proposal at a planning meeting, it is more than likely that he will be drawing upon his general knowledge of the area. Most of the qualitative data on social justice in Box 5.1 is based on such general knowledge.

The value of this sort of information is limited by the fact that it tends to be qualitative rather than quantitative and to have a high degree of subjectivity. Furthermore, its coverage is seldom comprehensive or systematic, since it has been acquired coincidentally, rather than through a planned data collection exercise. However, these disadvantages are at least partially counterbalanced by the depth and vividness which such information inevitably has. In many cases, it is possible to extract and utilize this sort of information in a more systematic manner; for example, relevant informants may be asked to complete a questionnaire or attend a structured interview. However, in this case one is, strictly speaking, moving from secondary to primary data collection. This sort of primary data collection will be discussed in the last part of section 5.4.

5.4 The collection and use of primary data

Primary data is, as already indicated, data which is collected for a specific planning purpose. In most kinds of planning activity it is necessary to collect some primary data because secondary data seldom provides all the information needed. This is perhaps particularly true in the case of social planning, since secondary data is unlikely to provide the detailed information about people's needs, problems and attitudes which such planning frequently requires. However, there are several different ways of collecting primary data for social planning purposes, depending partly on the type of information required but also on the resources available for data collection. Although primary data collection generally requires more time, money and manpower than using secondary data, there are methods which are relatively inexpensive. This section of the chapter looks at four different types of primary data collection, in order to demonstrate the variety of approaches possible and the issues and problems likely to arise. The four types are: censuses; sample household surveys; rapid appraisal techniques; and participant observation.

Censuses

The term 'census' is generally used to refer to a survey of all the households in a given area in order to obtain basic data on household characteristics, particularly (but not only) demographic characteristics. The most obvious examples are national population censuses, which were discussed in section 5.3 under the heading of demographic data. It was suggested in section 5.3 that conventional censuses tend to be particularly expensive forms of data collection, since they involve visits to every household in the area. Very few districts ever have the resources to undertake this sort of census for the district as a whole. There are, however, two types of census which can be considered at district level.

'Indirect' census

One is what might be called an 'indirect' census. In this case, household information is obtained indirectly from informants, rather than by actually visiting each household. The informants are usually local level officials, who are delegated the task of visiting each household in the areas under their jurisdiction and obtaining certain prescribed information from them. This approach only works if the information required is kept very simple and there is an effective network of local officials who can be relied upon to undertake the work with a reasonable degree of accuracy and integrity. It is most effective if the officials themselves, and preferably also the people from whom they have to collect the information, appreciate the purpose and value of the exercise. And even then one cannot expect the information to be highly accurate.

This sort of census is worth trying if one requires basic, and not highly accurate, information on the population of the district as a whole (or a large portion of it) and there is no appropriate secondary source. It is particularly useful if there is no reasonably accurate or up-to-date information on population size, or if such information is not disaggregated on the basis of administrative units which are generally used for district planning purposes. Box 5.4 describes a census which was undertaken in Gondwanaland District in 1988, in order to determine the approximate population of each ward and village as a basis for distributing the district's annual Rural Development Fund allocation. In this case, the system of ward and village development committees was used as the basis for data collection.

BOX 5.4

THE 1988 GONDWANALAND CENSUS

In 1988 the Gondwanaland District Development Committee decided that there was a need to know the approximate population of each ward, as a basis for determining the distribution of the annual Rural Development Fund (RDF) allocation between wards and for general planning purposes. Since it was eight years since the last national population census, there was no up-to-date population data The next census was not due until 1990 and the results of this would probably not be available before 1992. They therefore decided to undertake their own census, using the ward and village development committee structures. The results would not be as accurate as a proper census, but it would be better than nothing. The District Secretary (DS) and the Chief Executive Officer (CEO) of the District Council were given the responsibility of organizing the census.

The census was undertaken in the following way:

    · At the next meeting of the District Council, the DS and CEO explained that it was necessary to know how many people there were in each ward, in order to allocate the RDF fairly and to know how many people were supposed to pay poll tax. The point about tax was introduced to discourage people from over-estimating the population in order to gain a larger share of the RDF.

    · The councillors, who chair the ward development committees, were given a census form for each village in their ward. On this form they were required to list all households in the village and to record the name of the household head and the number of people in each. They were briefed on how to complete the forms at a special meeting, which was also attended by the ward community development workers (CDWs).

    · Each councillor, assisted by the CDW, then held a meeting of his (or her) ward development committee, which is composed of the chairpersons of each village development committee. At this meeting the census was explained and a form was given to each village chairperson.

    · The village chairpersons then conducted the census, monitored as far as possible by the councillor and the CDW.

    · The forms were then returned to the councillor and from there to the Council, where they were checked and processed by the CEO and his staff.

Census of a small area

The other type of census which can be attempted at district level is one which involves only a very small area or community. This kind of census can be very useful if one is planning a project in a particular community (such as a village) and requires basic information on all households in the community in order to design the details of the project and/or provide 'baseline' data from which to later evaluate the impact of the project. In such cases, special funds are sometimes available for such preparatory studies, especially if the project is being funded by some sort of donor agency. And if special funds are not available, existing resources (such as extension staff) can usually be diverted from their normal duties for the relatively short period of time needed to undertake the census.

Box 5.5 illustrates the use of such a census. It describes how the District Agricultural Officer in Gondwanaland decided, after the unfortunate experience with the vegetable garden project described in Box 3.3, to plan the next project of this kind more carefully. This was a project designed to increase cotton production in Zone IIb and the DAO decided that the first step should be to find out more about existing farming systems in the area. In order to do this he used various survey methods, the first of which (described in section (a)), was a census of two villages, in order to find out how much cotton was actually being grown and to provide a basis for selecting a smaller sample of farmers for more detailed study.

Sample household surveys

Sample household surveys are the best known method of obtaining information at the household level. Since such surveys include only a sample of the total population, they require less resources than a conventional census, and so are more likely to be a practicable means of obtaining information for district planning. However, they still require relatively large amounts of time, money and manpower. Consequently, they are generally used for specific purposes in a limited geographical area, rather than to provide basic data for the district as a whole.

BOX 5.5

COLLECTING INFORMATION FOR THE COTTON DEVELOPMENT PROJECT

In 1991 the District Agricultural Officer (DAO) of Gondwanaland decided to look at ways of increasing cotton production in Zone IIb, where the natural resources are similar to Zone IIa but cotton production was considerably lower. After the experience with the vegetable garden project in Zone III (see Box 3.3), he decided that rather than just promote cotton through normal extension methods, he would first find out more about farming systems in the area, in order to see if there were any particular constraints to cotton growing. Since he could not get information from every village, he selected five villages at varying distances from the only cotton marketing depot in the area. He then proceeded to conduct three different kinds of survey: village censuses, sample household surveys and rapid rural appraisal.

· Village censuses

His original intention was to conduct a complete household census in each of the five villages, in order to find out how much cotton each was growing as a basis for selecting a smaller number of households for more detailed study. However, since his resources were limited, he decided to begin in two villages. The censuses were conducted by one of his staff with some training in survey methods, assisted in each case by the extension worker responsible for the village. They went to the villages during the cotton growing season and made rough estimates of the area under cotton planted by each household.

· Sample household surveys

The DAO then divided the population of each of the two villages into three categories on the basis of the amount of cotton grown. The three categories were: none, up to one hectare, and more than one hectare. 10% of the households in each category were then selected randomly for more detailed study. In this case, a simple questionnaire was used, containing questions on household size and composition, ethnolinguistic origin, total area cultivated, area under cotton, availability of draught power, and distance from the cotton marketing depot. The findings from these two surveys suggested that there were three main factors affecting the amount of cotton grown: ethnolinguistic origin (Gonds appearing more likely to cultivate cotton than Wana), availability of draught power, and distance from the depot.

· Rapid rural appraisal

The DAO was pleased with the results of the studies in the first two villages but concerned about the amount of time they had taken. He therefore decided in the other three villages to replace the census with a rapid rural appraisal. Instead of visiting each household, the two extension workers sat down with the village leaders, and together they drew a sketchmap showing the location of all the households in the village, discussed how much cotton each household grew and divided them into the same three categories used in the other villages. They then cross-checked this information by visiting a few households in each category. The sample household surveys were then conducted in the same manner as in the first two villages.

The sample can be selected in three main ways, depending on the type of information required:

In all cases the size of the sample is important, since the sample must be large enough to provide data which is statistically representative of the population as a whole. The minimum sample size varies according to the size of the population and the type of sample. For example, the proportion of the population included in the sample should be larger in small than in large populations, and random samples should be larger than stratified random or purposive samples. However, it should not normally be less than about 10% of the population.

Since such surveys are a well established method of social research, on which much has already been written, further discussion of their methodology is unnecessary here. However, Box 5.5 provides an example of the use of a sample household survey. It describes how the information on cotton production obtained from the two village censuses in Zone IIb of Gondwanaland was used as a basis for selecting a stratified random sample of households in each village, who were then interviewed in more detail. In this case, the sample was stratified on the basis of the amount of cotton grown and the main purpose of the sample survey was to find out more about the factors affecting cotton production.

Rapid appraisal techniques

The concept of 'rapid rural appraisal' was introduced in the late 1970s as a means of obtaining information for rural development planning in situations where there is insufficient time, money and/or manpower to use conventional social survey methods, such as censuses and sample surveys. Until that time, the only alternative source of information was casual observation, often based on a quick drive around the area concerned and thus referred to as 'rural development tourism', which provided unreliable data, biased in favour of those areas easily accessible by road.

Since then, the term 'rapid rural appraisal' has become widely used and, in some cases, extended to cover more than just primary data collection. Because the techniques it uses to collect data often involve active participation by local people, it has in particular become closely associated with participatory planning and the term is thus sometimes used to refer to a comprehensive participatory planning approach. This dimension of rapid rural appraisal will he examined in Chapter 9, which looks specifically at participatory planning. However, for the purposes of this chapter the term will be used in its original sense of a means of collecting primary data in situations where there are insufficient resources to adopt conventional survey methods.

The basic aim of rapid rural appraisal is to obtain information which is adequate - in terms of both quantity and quality - to achieve a particular planning objective with the limited resources of time, money and manpower available. Because the techniques used are in a sense 'short-cut' methods of data collection, the information obtained is often not as accurate or comprehensive as that which would be obtained from a conventional census or sample survey, and it tends to be more qualitative and subjective in form. However, this is not necessarily the case. Much depends on how well the two types of survey are conducted. The results of a good rapid rural appraisal are likely to be much more accurate than those of a bad census or sample survey. Moreover, because rapid appraisal techniques tend to require the active participation of the people who will benefit from the project or programme being planned, there are additional benefits (which will be discussed in Chapter 9) in terms of the long-term sustainability of the project or programme.

Rapid rural appraisal is an art rather than a science, in that it involves designing data collection programmes to fit the particular needs and constraints of each situation, rather than rigidly applying predetermined survey methods. Consequently, there are many different types of appraisal techniques and new ones are being 'invented' all the time. Furthermore, in many cases, more than one technique is used to obtain the same information, as a means of cross-checking the information obtained and thereby compensating for any inaccuracies or biases.

Because there are so many different types of rapid appraisal technique, it is not possible to even list them all here, let alone describe them in any detail. The best way of appreciating their nature and scope is to look at specific examples. Boxes 5.5 and 5.6 provide two such examples. Section (c) of Box 5.5 is a continuation of the account of the cotton expansion project. It describes how the DAO extended the survey to three more villages but, in order to reduce the resources required, used rapid appraisal techniques instead of a full village census to find out how much cotton was being grown and provide the basis for the stratified random sampling. And Box 5.6 describes how various rapid appraisal techniques were used to obtain basic information about farming systems throughout the district, in order to supplement the limited information available from the National Sample Agricultural Survey. The 'indirect' census described in Box 5.4 might also be described as a form of rapid appraisal.

Participant observation

The term "participant observation" is used to refer to the collection of data by people who are actively involved in a situation or an area in a capacity other than data collection. Such people play a dual role, in that they collect data in addition to their normal daily functions. Consequently, participant observation is a relatively economical means of data collection and is, therefore, often included among the repertoire of rapid rural appraisal techniques. However, it is discussed separately here, partly because it is a well established means of data collection which was used before the concept of rapid rural appraisal was conceived and partly because it is of particular relevance in district planning.

Its relevance to district planning lies in the fact that extension workers (including not only agricultural staff but also community development workers, teachers, health workers, and so on) are a valuable potential resource of participant observers. It was mentioned in the last part of section 5.3 that such staff have a great deal of general knowledge about the district which is seldom used systematically but can be 'harnessed' for planning purposes.

BOX 5.6

USING RAPID RURAL APPRAISAL TO GET FARMING SYSTEMS DATA

Encouraged by the findings of his surveys of cotton production in Zone IIb (see Box 5.5), the District Agricultural Officer (DAO) of Gondwanaland decided to use rapid rural appraisal methods to obtain basic information on the farming systems in each zone. This would provide more detailed information than that available from the 1991 National Sample Agricultural Survey and would be useful for basic agricultural planning purposes.

His aim was to obtain the following information for each agroeconomic zone:

    1. Estimated area cultivated (average and range)

    2. Main crops cultivated (cash and subsistence)

    3. Estimated proportion of households using:

    (a) irrigation

    (b) draught power (ox/tractor)

    (c) fertilizers (mulch/manure/artificial)

    (d) hired labour

    4. Estimated number of livestock owned, by type (average and range)

    5. Main uses of livestock (meat, milk, draught power, cash sale, social status)

    6. Grazing system (communal/paddocks/stall-feeding).

He realized that he would not be able to get highly accurate information, especially in the case of the quantitative data But he hoped to be able to get enough information to give a general picture of the farming systems in each zone.

The data collection involved five main stages:

· Questionnaires covering the above topics were sent to all agricultural extension staff They were required to complete a questionnaire for the ward in which they were working (and/or any other wards with which they were familiar if they had been transfered recently), using their existing general knowledge of the area rather than conducting any kind of special investigation.

· The questionnaires were analyzed by zone. The information was adequate to prepare a general description of each zone and, in some cases, to divide a zone into 'sub-zones' if there appeared to be significant differences within it. However, the quantitative data on area cultivated and livestock numbers was inadequate to draw any meaningful conclusions.

· In order to cross-check the initial information and get more accurate quantitative data, the following steps were taken:

    (i) Extension staff were required to cross-check their information on area cultivated and livestock numbers by asking specific questions to those farmers whom they visited in the course of their daily work.

    (ii) A small team of extension staff visited a small sample of villages in each zone and sub-zone (selected randomly) and cross-checked the information through observation and discussions with village leaders.

    (iii) Comparisons were made with any available secondary data

· This information was used to revise and supplement the initial descriptions of zones and sub-zones.

· Finally, the DAO organized a workshop, attended by all extension staff at which the descriptions of zones and sub-zones were discussed and final amendments made.

There are two main ways of harnessing this knowledge. One way is to extract information which the extension workers already have. This can be done by, for example, requiring them to complete a questionnaire, attend a structured interview or take part in a roundtable discussion. The other way is to ask them in advance to collect specific information in the course of their daily work and to record their observations in a particular form.

Both methods are illustrated in Box 5.6. The agricultural extension workers of Gondwanaland were required, initially, to complete a simple questionnaire about the farming systems in the agroeconomic zones with which they were familiar and, subsequently, as an extension of their normal duties to collect certain information which was then used to check and supplement the initial data.

SUMMARY

RECOMMENDED READING

Casley, D.J. & D.A. Lury, Data Collection in Developing Countries, Oxford, Clarendon Press, 1981. A comprehensive and practical guide to many forms of data collection required at district level, although with the emphasis on conventional rather than 'rapid' techniques.

FAO, Guide for Training in the Formulation of Agricultural and Rural Investment Projects: Planning Tools, FAO, Rome, 1994. A training text with exercises on survey methods and rapid rural appraisal.

FAO, Population, Society and Agricultural Planning, FAO Economic and Social Development Paper 51, Rome, 1987. Includes detailed guidelines on the collection and analysis of demographic data and its application in various aspects of agricultural planning.

Longhurst, R. (ed), 'Rapid rural appraisal', Bulletin of the Institute of Development Studies (university of Sussex), vol. 12, no. 4, October 1981 (whole issue). The first comprehensive collection of articles on rapid rural appraisal; indicates its rationale and the nature and scope of techniques used.

McCracken, J.A., J.N. Pretty & G.R. Conway, An Introduction to Rapid Rural Appraisal for Agricultural Development, London, International Institute for Environment and Development, 1988. Useful overview of the objectives and methods of rapid rural appraisal and their application for agricultural planning; emphasizes the role of participatory techniques.

Peil, M. et al., Social Science Research Methods: An Africa Handbook, London, Hodder & Stoughton, 1982. A comprehensive and practical review of conventional social survey methods; written specifically for Africa but relevant to most developing countries.

Chapter 6 Formulating social policy

This chapter looks at the methodological issues and problems which arise when formulating policies to achieve objectives which are primarily or entirely 'social' in nature. It does this by taking three specific examples: nutrition planning; education planning; and planning rural water supplies. In each case, four aspects of planning are considered: its role in the overall process of rural development; data needs; major policy issues; and organizational implications.

6.1 Introduction

This chapter is concerned with the formulation of policies to meet social needs or objectives. Its aim is to indicate the types of methodological issues and problems which arise at the policy formulation stage. Chapter 7 will then look at the next stage of the planning process, that of planning specific projects and programmes. In most countries the main responsibility for policy formulation rests at national level and the scope for making major policy decisions at district level is thus limited. However, this does not mean that there is no role at all for policy formulation at district level. This chapter concentrates on those aspects of policy which can be planned at this level, within the broader context of national policies.

The focus of the chapter is on planning to achieve objectives which are generally regarded as being primarily or entirely 'social' in nature, rather than planning where the main objectives are, say, economic or environmental and social considerations are thus of secondary importance. However, this does not mean that 'non-social' factors can be ignored. As already emphasized in Part I of these Guidelines, social development must be seen as an integral part of the overall process of rural development and so cannot be planned in isolation. Consequently, all the policy issues discussed in this chapter have economic as well as social implications and require inputs from more than one discipline or government department.

Since the term 'social' is very broad, there are many different kinds of social policy, each of which raises its own issues and problems. Therefore, instead of trying to generalize about social policy formulation as a whole, the chapter examines three specific examples: nutrition, education and rural water supply. However, in order to emphasize those issues and problems common to all social policy formulation, the discussion of each will be organized under four sub-headings: its role in the overall process of rural development; data needs; major policy issues; and organizational implications. And in each case, the points made will be illustrated by references to relevant sections of the 1992-97 Gondwanaland District Five-Year Plan, which was produced by the District Development Committee in 1991 as a basis for district planning.

6.2 Nutrition policy

Nutrition is not only an important focus of attention in its own right but also a major component of other aspects of rural development, especially agriculture and health. This is reflected in the fact that there is seldom a separate government agency responsible solely for nutrition at national level. It is thus a good example of the inter-disciplinary nature of social policy formulation.

The role of nutrition in rural development

Nutrition both affects and is affected by other aspects of rural development. In other words, it may be regarded as both an input into the development process and an outcome of it.

It is an input to development in two different ways. On the one hand, it is an important objective in itself, in that an improvement in nutrition increases the general quality of life of the persons concerned. And on the other hand, it is a means of achieving other objectives, notably better health and, therefore, increased productivity, reductions in the cost of health care to the individual and the state, and (again) an increase in the general quality of life.

Nutrition is an outcome of development in that an individual or community's nutritional position (or 'status') depends on many different factors, all of which can be influenced directly or indirectly by development activities. These factors include the quantity and quality of food available, household income and expenditure preferences, family size and (in the case of infants) child spacing, gender relations in both the production and consumption of food, and knowledge about nutrition. It is therefore affected by both the general level of development of the individual or community and by specific development programmes, especially in the fields of agriculture, education and community development.

However, the relationship between nutrition and other aspects of development is often complex. One example of this complexity, that of the impact of cash cropping on nutrition, was given in Chapter 3 (section 3.3) and illustrated by the example of the Gondwanaland vegetable garden project (Box 3.3). Another example is the relationship between income and nutrition. An increase in individual or household income generally results in an increase in the amount of food purchased and consumed, at least up to a certain income level at which nutritional needs are met. But it does not necessarily result in an improvement in the quality of food consumed. In fact, it may result in the consumption of more expensive but less nutritious foods - for example, white rice instead of brown rice, processed instead of unprocessed foods, mineral drinks instead of water or tea, and increased amounts of sugar, fats, alcohol and so on. Moreover, it may also result in the consumption of too much food, resulting in obesity, high blood pressure, heart disease and related problems, especially when accompanied by a reduction in exercise and/or in cultures where food consumption is regarded as a symbol of affluence and social status. The implications of these complexities for nutritional policy will be considered later.

Data needs

In order to formulate policies to improve nutrition there is a need for information on both the nutritional status of the area or community concerned and the factors affecting this status. Unfortunately, neither kind of information is easy to obtain, especially at district level where resources are limited.

There are two main ways of measuring nutritional status. One is to measure the actual intake of food, in terms of both quantity and nutritional value. This is difficult to do without detailed household surveys, although general information can be obtained by rapid appraisal methods (see Chapter 5), such as discussions with a group of villagers (especially women) or observation by extension staff who live and work in the area. The other method is to measure the impact of nutrition on body size and/or weight. This is not a reliable way of measuring the nutritional status of adults, except in cases of severe malnutrition. However, it is a reasonably accurate indicator of nutrition in the case of small children and there are two relatively simple yardsticks available: weight in relation to age and the circumference of the upper arm. In most countries local medical personnel keep such records for all under-fives who attend maternal and child health clinics; these records are usually the most easily accessible - and often the only - secondary data on nutritional status available to planners at district level.

In order to get information on the factors affecting nutritional status, it is necessary to understand the whole process of food production and/or purchase, storage, preparation and consumption at household level. As in the case of food intake, this really requires detailed household surveys, although general information can be obtained from group interviews and casual observation. It is particularly important - and difficult - to obtain information on relevant cultural factors, such as food preferences, traditional methods of food preparation, and the division of both labour and food between men and women.

Major policy issues

There are many different policy issues related to nutrition. However, for purposes of rural development planning at district level, the following are of particular importance:

· Poverty:

There is a close relationship between poverty and malnutrition. In most rural areas, therefore, an increase in income (either in cash or in kind - ie. food) is a necessary, although not always a sufficient, requirement for improving nutritional status. Consequently, the alleviation of poverty must be a major part of any strategy for improving nutrition. And conversely, any 'development' strategy which results in increased poverty among some or all of the population is likely to have an adverse effect on nutrition. One of the main concerns about structural adjustment programmes is the increase in poverty and, therefore, malnutrition which tends to occur, due primarily to unemployment and increases in food prices. There is little that planners at district level can do to affect such strategies directly. But they can help to alleviate their effects by encouraging income-generating activities and they may be able to have an indirect effect on national policy by monitoring and publicising increases in poverty and malnutrition.

· Subsistence production:

There is a need in many countries to give more attention to subsistence production, with the aim of increasing the quantity, quality and reliability of food supply at household level. The strategies adopted will vary from place to place, since they must be compatible with local physical and cultural conditions. However, they are likely to include measures to increase yields, introduce more nutritious and/or reliable (eg. drought resistant) crops, and improve storage facilities. Agricultural staff at district level usually have some scope to initiate or give greater attention to such measures, although they are obviously bound by national policy and the resources available.

· Cash cropping:

Although cash crops are important as a means of increasing income and therefore reducing poverty, it is important that cash cropping is not promoted at the expense of nutritional considerations. For example, farmers should not be encouraged to devote so much land or inputs to cash crops that they cannot grow enough food to eat, unless it is certain the money earned from cash cropping will be enough to buy food - and that food is available to buy. From this point of view, food crops have an advantage over other cash crops, in that they can be used for domestic consumption and (if there is a surplus) for sale. This has implications in terms of extension policy at district level. It also has implications for pricing policies, but these are obviously beyond the control of district planners.

· Nutritional education:

The policies discussed so far will not result in significant improvements in nutrition unless they are accompanied by nutritional education. This can be provided through various means, including schools, clinics, community development workers, and agricultural extension staff. And it is something that can be initiated at district level, although again there will be constraints due to national policy (eg. school curriculum) and resource availability.

· Food relief:

Finally, there is a need to be prepared for the possibility of having to provide emergency food relief, particularly in areas susceptible to natural disasters such as drought or floods or affected by war or civil unrest. Although this is primarily a national responsibility, district staff can and should take some precautionary measures, especially if their areas are vulnerable. For example, they can look out for early signs of critical food shortages and have an emergency plan ready to put into action as and when needed. And they may also be able to establish emergency grain stores.

Box 6.1 describes how these various policy issues were incorporated into the 199297 Five-Year Plan for Gondwanaland District.

BOX 6. 1

NUTRITION POLICY IN GONDWANALAND

The Gondwanaland District Five-Year Plan for the period 1992-97 includes a policy statement on nutrition. The need for such a policy arose from concern by the Ministry of Health and Council health staff about the number of children suffering from malnutrition in the district and observations by Agriculture staff about the possible negative impact of cash cropping on nutrition. The latter included evidence from the vegetable garden project in Zone III (see Box 3.3), which demonstrated the complexity of the relationship between cash cropping and nutrition, and the survey of cotton production in Zone IIb (Box 5.5), which revealed a negative correlation between the amount of cotton grown and the amount of food crops, especially in the case of smaller-scale farmers. There was particular concern about the situation in Zone V, where there is a very high rate of malnutrition (compare Box 3.4), emergency food relief is often needed, and its distribution is frequently hampered because of the area's inaccessibility in the wet season.

The main components of the nutrition policy are:

    · There will be a concerted effort to alleviate poverty in the district, in order to (among other things) improve the nutritional status of the population.

    · The Ministry of Agriculture and Natural Resources will prepare a strategy for improving the production and storage of food crops for each agroeconomic zone and promote these strategies through its extension work. Special attention will be given to Zone V, because of the particular problems there.

    · Before embarking on any attempt to either increase production of an existing cash crop or introduce a new one, the likely implications in terms of nutrition will be assessed.

    · A simple package of materials on the basic principles of nutrition will be prepared by the Ministry of Health, in consultation with other relevant agencies, for use by schools, clinics, community development workers and agricultural extension staff

    · Agriculture staff will provide advance warning of any likely food shortages to the district's Disaster Relief Committee, which win then be responsible for organizing the necessary food relief A special store of food grain will be kept at the clinic in Zone V, to facilitate the distribution of relief food when required.

The policy was formulated by a Nutrition Task Force, composed of representatives of the Ministries of Agriculture and Natural Resources, Health, Education, and Community Development and Social Welfare, and the District Council. This Task Force is also responsible for coordinating and monitoring the implementation of the policy.

Organizational implications

There are two main implications in terms of the organizational structures and procedures needed to formulate nutrition policy at district level:

· The formulation of nutrition policy must be an inter-departmental activity, since several different government departments or agencies are involved. The most obvious are those responsible for agriculture, health, education and community development. This suggests the need for an interdepartmental committee or task force at district level, like that established in Gondwanaland (see Box 6.1). If it is necessary to appoint one agency to play a lead role (for example, to chair the committee), agriculture or health is likely to be the most appropriate; however, this will depend partly on where the main responsibility for nutrition lies at national level.

· Ultimately, decisions about nutrition are made at individual or household level. Consequently, it is essential that the needs and priorities of local people and the social structure within the household are understood and taken into consideration in policy formulation. This suggests the need for a participatory approach.

6.3 Education policy

Education is a more obviously 'sectoral' activity than nutrition, in the sense that responsibility for education is usually clearly located within one central government ministry or department, although some educational functions are often delegated to local governments. It thus illustrates the process of policy formulation within one of the main 'social sectors'.

The role of education in rural development

Like nutrition, education may be regarded as both an input into the development process and an outcome of it. As an input, it is - again like nutrition - both an objective in its own right and a means of achieving other objectives, notably the establishment of a skilled labour force and the creation of a generally educated and aware population, which in turn result in increased productivity, a reduction in the birth rate and therefore in the rate of population growth, increased awareness of human rights and responsibilities, a more enlightened electorate, and so on. And it is an outcome in the sense that both the quantity and quality of education available to an individual or a community depends on other factors, particularly the affluence of individual households and communities and of the nation as a whole, 'which is in turn dependent on the general level of economic development.

However, the relationship between education and other aspects of development is also complex. For example, education is not always beneficial to the general development of a nation. Its value depends on the extent to which the quantity, type and quality of education matches development needs. Many less developed countries have large numbers of unemployed school leavers who are reluctant to work on the land but cannot find non-agricultural employment, while at the same time industries, professions and government suffer from a lack of competent and/or qualified skilled manpower. The implications of this for agriculture were discussed in Chapter 3 (section 3.3) and illustrated by the example of the Gondwanaland vegetable garden project (Box 3.3). In other cases, the problem is that the education system is primarily a means of indoctrination rather than education, and so does little to create an educated and aware population.

Similarly, although access to education usually increases if there is an increase in wealth at household, community or national level, this is not always the case, since it also depends on the relative importance attached to education by the household, community or nation. For example, attitudes to education in general, and the education of girls in particular, vary significantly from one household to another and from one culture to another. And the proportion of national income allocated to education varies considerably from one country to another.

Data needs

In order to formulate education policy at district level, information is needed on both the supply of and demand for education. In other words, there is a need, on the one hand, to have data on existing and potential education facilities and, on the other hand, to know the number of people requiring education and the type of education needed.

The supply side data is generally relatively easy to obtain. As indicated in Chapter 5, district education offices usually keep reasonably good records, including information on the number and location of schools, the quality of buildings, the number of pupils by school, class and sex, the number of teachers and therefore the pupil : teacher ratio, school curriculum, performance in examinations, and so on. Moreover, education staff generally have some idea of the resources likely to be available for future expansion.

Data on demand tends to be somewhat more complex and more difficult to obtain. There are three main kinds of information needed.

Firstly, there is a need for quantitative data on the numbers of children eligible to attend various types or levels of education. For example, the number eligible to attend primary school can be determined by calculating the number of children of primary school age (see Box 5.3 for an example), while the number eligible to attend secondary school can be calculated on the basis of the number completing primary school or (if entry to secondary school is restricted) meeting the necessary entrance requirements.

Secondly, there is a need for data on the geographical distribution of demand. This can best be obtained by locating educational facilities on a map and relating this to the distribution of population. Box 6.2 shows the distribution of population by village in Gondwanaland District (based on data collected in the District Census described in Box 5.4) and the location of secondary schools. The circles around the schools indicate their catchment areas, based on a policy that children should not have to travel more than 10 kilometres to secondary school. This map can therefore be used to calculate the proportion of the population living within 10 kilometres of a secondary school (which was one of the social indicators included in Boxes 3.2 and 5.1) and to indicate those parts of the district where the provision of secondary schools is inadequate because children have to travel more than 10 kilometres and, if necessary, the approximate numbers of people thus deprived. If detailed population data is not available, the numbers of people who do or do not live within the stipulated distance cannot be calculated, but the catchment 'circles' can still be used to indicate the geographical areas where provision is inadequate.

Thirdly, there is a need for information on the demand for skilled or educated manpower, in order to try to match education with manpower needs. This is the most difficult kind of data to obtain, especially at district level, since school leavers are likely to look for jobs outside the district as well as within it. Moreover, district planners may feel that there is little point in collecting such information if (as is usually the case) they have no power to either restrict school entry or change the curriculum on the basis of manpower needs. Nevertheless, it is generally possible - and useful - to obtain some information, albeit of a qualitative nature, on the scale of unemployment among school leavers in the district, the types of skills required by any major employers in the area, and the relevance of the curriculum to smallholder agricultural production and small-scale or informal non-agricultural income-earning opportunities.

Major policy issues

Once again there are many different policy issues which might be considered. However, since the aim here is to illustrate the types of issues involved rather than provide a comprehensive guide to educational policy, attention will be focused on four main policy areas:

Box 6.2 Gondwanaland: access to secondary schools

· The level of education:

Education is usually categorised on the basis of 'level', the main levels being pre-school, primary, secondary, tertiary and adult. Changes in the amount of resources devoted to any or all of these affect the impact of education on other aspects of development. For example, primary and adult education are seen primarily as a means of creating an educated and aware population, while secondary and tertiary education are directed more towards meeting the need for skilled manpower. The scope for district planners to influence the relative importance of these various levels of education is limited, since such decisions are generally considered part of national policy. However, they may be able to have some impact, particularly if they have some control over resources. For example, if they have control over funds to support the construction of school buildings, they can decide to, say, give higher priority to primary than secondary schools. Similarly, if there is a general rural development fund, this can used to support adult education.

· Access to education:

The total number of people in