Questionnaires and instruction manuals
Since it is not the main objective of a general-purpose agricultural production census or sample survey to study in detail the structure of the population in agriculture, the questionnaire items should be limited in this respect. The FAO Programme for the World Census of Agriculture 2000 includes specific proposals for agricultural census questionnaire items (FAO, 1995).
As a general rule, data on the agriculture-related population should be collected through the regular agricultural surveys only if such data cannot be obtained through the population census, through a periodic population sample survey or through a special agricultural sample survey. In other words, the selection of the agricultural survey questionnaire items on the population concerned with agriculture should always consider and be a function of the following factors:
· The availability of such data from the population and housing census or periodic population sample surveys.
· The possibility of including these items in a population survey or census questionnaire.
As has already been mentioned, it is important that countries coordinate adequately their agricultural census and survey programmes with the population and housing census and survey programmes which are generally organized by different government institutions.
The language used in the survey questionnaires and in the instruction manuals should be as simple as possible. Words that are not common to the agricultural population of the country should not be used. The use of uncommon words or expressions might place female respondents at a disadvantage since women are generally less exposed to public contact and training than men and frequently have a lower level of education.
In particular, the word "gender", or other words that are not common to the agricultural population in developing countries, should not be used in the questionnaires and instruction manuals in place of the word "sex". In fact, they are for use by farmers or enumerators in developing countries, a large percentage of whom may not know the word. In the future, if "gender" becomes universally known, then it may be used, as is now recommended by some data users and analysts.
The use of words or expressions in the official language that may be gender biased should be avoided.
Special attention should be given to the translation of the questionnaires and manuals from the official language into the spoken dialect. Words that are not gender biased in the official language may become biased in the translation.
The way questions are asked affects the answers. This is particularly true for women respondents and for questions regarding economic activities. Studies on this problem have shown that by providing examples of what is meant by "activities", women's work is more likely to be captured. The order in which questions are asked also affects the answers, particularly on gender issues, for several reasons - for example, respondent fatigue.
Instruction manuals should be adequately balanced with male and female examples and illustrations.
Qualified staff of both sexes on the subject matter should be available during the preparation of the instruction manuals. Gender neutrality is very important and should be maintained while preparing the manuals.
