Integration of gender concerns into the planning and design of agricultural censuses and surveys
The adoption of a gender perspective in agricultural censuses or surveys has to be decided in the first stages of planning of the agricultural census or survey, because it has significant implications in terms of topics covered, census and surveys operations such as data collection design and training of field staff, and, at a later stage, data analysis and dissemination. An analysis of agricultural censuses undertaken in Africa in the 2000 census round showed, for example, that the production of statistics on gender and agriculture is improved when the need for gender statistics is incorporated into the objectives and scope of the censuses (FAO Regional Office for Africa, 2005). The analysis also showed that the process of obtaining gender statistics improves when potential users of gender statistics, with clear demands of specific data, are involved in the preparation of the censuses.
Coverage of all relevant units of enumeration
Gender bias in data collection can be introduced by improper coverage of all relevant units of enumeration. The unit of enumeration in agricultural censuses and surveys is the agricultural holding. An agricultural holding is an economic unit of agricultural production under single management, comprising all livestock kept and all land used wholly or partly for agricultural production purposes, without regard to title, legal form or size (FAO, 2007). There are two types of agricultural holdings: (a) holdings in the household sector – that is, those operated by household members; and (b) holdings in the non-household sector, such as corporations and government institutions.
Proper coverage of the household sector is the most important from the perspective of generating gender statistics. The exclusion of small holdings, a sub-sector where women and family members play a particularly important role, can be a drawback of agricultural censuses or surveys. When holdings below a certain size and/or holdings located in urban or peri-urban areas are excluded from censuses and surveys, women’s contribution to agricultural production may be underestimated. In addition, these excluded holdings could be playing an important role in food production and food security.
The inclusion of all units relevant to agricultural production needs to be considered when preparing the frame of agricultural holdings in the household sector and when designing the sampling frame for agricultural surveys. In particular, frames based on administrative sources where women are less likely than men to have their holdings registered can introduce significant gender-bias in data obtained. Gender differentials also need to be considered when deciding the stratification variables in the sampling design of agricultural surveys. When necessary, oversampling in one or more strata should be considered to allow for an adequate number of both women- and men-operated holdings in each stratum.
Adequate units of data collection and analysis
Comprehensive coverage of gender issues in agricultural censuses and surveys requires the use of multiple units both in the data collection stage as well as in the data analysis stage. As noted before, the unit of enumeration in agricultural censuses and surveys is the agricultural holding. Many censuses and surveys collect most of the data at the level of agricultural holding. For example, data on inputs for agricultural production, such as seeds or pesticides, are usually purchased for the whole holding and therefore they are collected at the level of the holding. Other examples of items usually collected at the holding level are the use of irrigation or use of agricultural machinery.
However, it may be necessary to collect and analyse data at intra-holding level in order to get a true picture on gender issues. Data on farm labour, especially data on participation of household members in activities of agricultural production, can be better captured at individual level, along with data on sex, age, marital status, educational attainment, or other characteristics related to the type and amount of work performed on the farm or for other business.
Depending on the country, data on land use and livestock are sometimes more suitable to be collected for smaller units within the holding. For example, data on land use are often collected at the level of parcels or plots that compose a holding. More general, data can be collected at the sub-holding level. A sub-holding is defined as a single agricultural activity or group of activities managed by a particular person or group of persons (sub-holders) in the holder's household on behalf of the agricultural holder. There may be one or more sub-holdings in a holding. A sub-holding could comprise a single plot, a whole field, a whole parcel, or even the whole holding. A sub-holding could also be a livestock operation associated with a plot, field or parcel, or a livestock operation without any land.
The collection of data on crops and livestock at the more disaggregated level may be preferred for several reasons. First, collecting data at the level of parcels or plots reduces the errors in reporting. This is especially the case when agricultural households work on several different plots of land, and different individuals are in charge of each plot or crop; or, when some household members are responsible for herds that are separated from those of the main holder. Gathering the data at plot level and herd level may appear as more time consuming. However, when data are collected at the whole farm level, the respondents may have by themselves to add up the information on different plots to come up with the required answer, increasing the chances for non sampling errors like reporting errors. The quality of data improves when women and men in charge of each plot or crop respond separately to questions about the plots or crops for which they are responsible. The operator of each plot is more likely to know specific details about the size and quality of the plot, how much time each household member has spent working on various tasks on that particular plot. Second, plots from the same holding may differ in terms of land quality, the degree of land degradation and erosion, and the data collected at their level may explain differences in agricultural production. Finally, disaggregated data at the level of sub-holdings are crucial for understanding gender roles and decision-making within the agricultural holding.
It is, however, to be noted that in most cases the application of sub-holding concept on the ground poses many practical challenges and increases the cost of data collection. Before deciding to apply this concept for large scale survey like an agricultural census a careful evaluation of social customs and cost-benefits of using this concept is recommended.
Gender-specific conceptual and measurement issues related to the topics covered in agricultural censuses and surveys have to be adequately reflected in the design of questionnaires used (see box 3.3 for a checklist of the main points that should be taken into account in designing questionnaires). It is important, from a gender perspective, that the questionnaire corresponding to the household sector is structured by the needed level of data collection. In that respect, different modules may be designed for different levels of data collection and/or different topics. For example, data collected at individual level may be covered by a module on demographic, social and economic characteristics of household members, including involvement in agricultural and non-agricultural economic activities on and off the holding; and a module on characteristics of non-family agricultural labour. Identification of the owners of agricultural resources and sub-holders may also be based at data collected at the level of individual household members.
Some other modules of the questionnaire, such as those on livestock or land use, may be designed to collect data at sub-holding level with the possibility of identification of the sub-holder. The implementation of the sub-holding/sub-holder concepts, important from a gender perspective, may be complex. The approach used by a country will depend on national agricultural practices and social and cultural conditions, taking into consideration the data collection methodology already existing or suitable. For example, when countries have the practice of collecting data on crops at the plot level and data on livestock at the herd level, it is relatively straightforward to identify the women and men who are in charge of those parcels and herds (sub-holders). Alternatively, a smaller set of items can be collected at the level of sub-holdings separately from the main crop and livestock data, by asking specific questions about the type of crop and livestock activities carried out under the control of the sub-holder.
Box 3. 3 Incorporating a gender perspective in the design of questionnaires for agricultural censuses and surveys. A checklist
Members of the team designing the questionnaires have been trained in gender issues and gender-specific measurement issues related to family and non-family farm labour and role of women as managers of holdings and sub-holdings.
For the household sector, there is a clear indication of the items to be collected at holding level, at sub-holding level, or at individual level of household members and hired labourers. If possible, identification of holders, sub-holders and owners allow the link with the individual characteristics of the household members.
When identifying the sub-holders, use a series of questions about each household member to find out about the types of work each carried out on the holding and their role in managing agricultural production activities.
Use of a series of questions instead of one question to identify the household members owning by themselves or jointly with another person parcels/plots of land, livestock by type, and agricultural machinery.
Avoid language suggesting that holder or sub-holders are male.
When measuring economic activity, the question should have a note for the interviewers indicating the use of activity lists (provided in the manual) and follow up probing questions.
Language in the questionnaire should be carefully used to avoid that the agricultural work of women is perceived and reported as housework rather than as economic activity.
If household head needs to be identified, a short note on the questionnaire should indicate the criteria of identification
Selection and training of the field staff
The quality of data collected in agricultural censuses and surveys (similar to other data collection programmes) depends on the quality of staff selected and the training provided. Box 3.4 presents a list of factors that should be taken into account when incorporating a gender perspective in the preparation of manuals and training of interviewers for agricultural censuses and surveys.
It should be noted that men are greatly overrepresented among field staff in agricultural censuses and surveys, often as a result of using as interviewers workers in agricultural extension services who are predominantly men. In general, it is important that both women and men are selected as interviewers and supervisors, and both women and men are trained to obtain quality data from both female and male respondents. In particular, the recruitment of women operators should be seriously considered in countries where women farmers do not feel free to talk directly to male enumerators due to cultural factors.
Box 3. 4 Incorporating a gender perspective in the preparation of manuals and training of interviewers for agricultural censuses and surveys - A checklist
Identify the key gender issues prevailing in the agricultural sector in the country of interest.
Gender training should emphasize gender-related objectives and goals of the census.
Gender training should increase awareness about the role of women in managing holdings and sub-holdings.
Both women and men are selected as training instructors and as trainers presented in audio-visual materials.
Interviewers women and men should be trained to interview persons of the same sex and of the opposite sex.
The language and examples given in the manuals or training materials with regard to identification of agricultural holders and sub-holders and the household head are free of gender-based biases.
Manuals and training materials provide examples on identifying joint agricultural holders.
Manuals and training materials should show examples for identifying the real decision maker in the farm. In particular, persons who are usually absent from the household should not be declared as household head or agricultural holder.
Training provides guidelines in obtaining information from women and men in charge of each plot or crop.
When collecting data on economic activity, provide in the manuals lists of economic activities, including lists of own-account productive activities and probing lists, to avoid underreporting of women’s economic activity. Training should emphasize problems and stereotypes associated with women’s work.
The census advertising is an important tool for improving the census coverage, in particular, of small holdings managed by women, and the reporting of women’s agricultural activity. The presentations prepared for advertising should illustrate both women’s and men’s contribution to agricultural production. The choice of type of media should take into account the fact that women may have easier access to some types of media than others. For example, in certain groups of population, women are more likely than men to be illiterate. Women may be easier to reach by radio programmes targeted at women or by use of graphics in places where women tend to gather together.