3. Methodology
3.1. Study design
Semi-structured interviews alternated with discussion groups were chosen as research methods in order to accurately assess farmers’ knowledge at the individual and group level and to identify if differences in soil quality perception occur between farmers who use different production systems (for further explanations of this methodological rationale see [37]). All interviews and discussions were recorded on tape for analysis.
3.1.1. Initial individual interviews
Semi-structured interviews were used for gathering information on perceptions of soil quality indicators. These interviews took place at the farmer's house or in his field. Out of the 200 active rice farmers, 50 were chosen to be interviewed based on the following criteria: (a) farmers should own and work the fields themselves in order to minimize misunderstandings due to limited knowledge of the soil and the production systems; and (b) the three production systems had to be represented among the farmers. After contacting these potential farmers, only 32 of them were interviewed (3 using the conventional production system, 13 using the pre-germinated system, and 16 using the semi-direct system; Table 3) because not all farmers were prepared to spend the necessary time or were interested in participation in the study. This first round of individual interviews was held in November and December 2003. The farmers were presented two open and broad questions: (1) “What do you think is a good soil?” and (2) “How do you recognize a good soil?” This resulted in an inventory of farmers’ perceptions of soil quality and of the soil quality indicators they use.