Sources of data

DATA FROM POPULATION CENSUSES AND LABOUR FORCE SURVEYS

Information relating to the labour force presented in Tables 2, 3 and 4 is derived from data on the economically active population collected in population censuses and labour force surveys. These are, in fact, the most appropriate sources of information about the labour force and employment in general and the agricultural/non-agricultural differentials in labour force participation. However, as a result of measurement problems (discussed in Box 1, Annex 1), it is clear that the extent of women's participation in agricultural work can be significantly underestimated in these sources. Consequently, efforts should be made to improve measurement techniques, in order to quantify more accurately the extent of female participation in agricultural work. Furthermore, more accurate details regarding the different categories of agricultural labour, as well as the hours worked (including gender differentials), should be obtained from these sources through the appropriate retabulation of the results.

DATA FROM AGRICULTURAL CENSUSES

Agricultural censuses and surveys usually focus on the basic characteristics relating to production technologies and land use. The FAO Programme for the World Census of Agriculture 2000 (WCA 2000) has put special emphasis on various items that can provide gender-related structural information on agricultural activities. The Programme recommends that a comprehensive cross-tabulation of the following items be made by the sex of the landholder:

• General characteristics (purpose of production).
• Demographic characteristics (size of the holder's household; holders and members of their households by age, education and marital status).
• Employment (holders and members of their households who are economically active by age, main occupation and type of work [permanent, occasional]; and holdings by number of permanent workers [members of the holder's household and hired workers]).
• Land and water (number and area of holdings; tenure of land; land use; the area of land cleared under shifting cultivation each year);
• Crops (major temporary crops, other temporary crops, major permanent crops, other permanent crops, use of high-yielding varieties, use of fertilizers, use of pesticides).
• Livestock (the number of each kind of livestock; the purpose of each kind of livestock).
• Machinery and equipment (number of stationary power-producing machinery by source of supply; use of other agricultural machinery by source of supply).
• Buildings and other structures (use of non-residential buildings by tenure; area and volume of non-residential buildings by purpose).
• Other activities (number and area of forest trees; fishery activities and other kinds of aquaculture installations).

The comprehensive structural data described above can provide not only additional information on the economic status of holders by sex, but also more insight into agricultural practices, in order to study whether efficiency and productivity are related to the sex of the holder. However, in using these data, it should be remembered that agricultural censuses are generally conducted using agricultural holdings as the unit of enumeration, so that the data listed above relate to both the members of the holder's household as well as to the individual holder. Following are some examples of the important areas that can be analysed with these items of data:

• Overall, these data can be used as a framework for conducting further sample surveys for extension support programmes; for drawing up research plans in the development of gender-specific agricultural technologies for specific crops and a specific holding size/mix (a combination of land and livestock) and that support secondary activities (that could for raise the living standards of agricultural households). The value of these data could be further enhanced by grouping into agroclimatic or agro-ecological zones.
• Data on the sex of holders are useful in determining the role of women in decision-making and the proportion of women who are heads of families.
• Data on holdings by the size of the holder's household, together with details of employment status, ownership of land and assets, provide a measure for household well-being. Such analyses may indicate the level of economic support provided by the head of household, as well as the management of human resources, and help to reveal the causes of gender-biased poverty in rural areas.
• Data on the educational status of holders and members of their households can be related to the use of agricultural inputs and machinery and thus provide the basis for designing suitable extension plans to develop the potential of women in adopting the latest techniques or creating their own enterprises.

However, to be useful for both users and analysts, the information collected needs to be tabulated and presented in a form that emphasizes the relevant gender issues.

DATA FROM HOUSEHOLD SURVEYS

Since land productivity and the extent to which land can provide income for holders are limited, other (non-agricultural) sources of income are necessary to improve the status of female workers. Generally, agricultural censuses and surveys do not collect comprehensive data on the non-agricultural activities of agricultural households. Thus, household surveys are essential to address such issues for the purpose of agricultural and rural development. Most of the data normally required for relevant programmes are collected in household income/expenditure surveys. The information refers to the details of income and expenditure classified by the sex and economic activity of the head of household. However, special tabulations need to be undertaken.

The retabulation of data can provide details for compiling agricultural household sector accounts, classified by the sex of the head of household. These data can be used to study and compare the economic behaviour of agricultural households. Depending on the interest of the policy-makers, however, it should also be feasible to generate cross-tabulations as listed below to study various distributions relating to economic status:

• generation of income (agricultural, non-agricultural and total) by size of holding and number of livestock;
• generation of income (agricultural, non-agricultural and total) by size of household and sex of economically active population;
• generation of income (agricultural, non-agricultural and total) by number of economically active women in the household and holding size/asset class;
• input-output and capital-output ratios by size of holding and number of livestock.

AGRICULTURAL OUTPUT AND INCOME BY SEX

Reliable information on the contribution of men and women to production and consumption, as well as income generation in agriculture, cannot be obtained from the existing national censuses and surveys, largely because the unit of data collection for output and consumption is not an individual but holdings or households. To obtain such information, it is necessary to use time-use surveys, allowing for double measurement strategies of production and consumption variables - in time units as well as in value units, both at the individual and household levels. Time-use methods of data collection and measurement also make it possible to address productivity issues at the individual and household levels. Thus, when time-use data are added to data on the physical output of individuals and households, it becomes possible to determine whether a decrease in time input corresponds to a decrease in quantity or quality of individual or household output, or whether the time decrease shows that productivity of the individual or household has increased. Various methods for monetary valuation of household work are also available for approximating measurements (e.g. in orders of magnitude) of the income of informal family or farm workers from various routine household and farming activities. Indeed, in agriculture these activities are often dispersed in territory and time, which creates difficulties in recording them systematically.³These difficulties can be addressed by combinations of anthropological and statistical methods of data collection, providing qualitative indicators in addition to quantitative measures and methods of analysis (FAO, 1996).

³ For a more detailed review of these issues, see FAO (1978).