A semi-structured questionnaire provided opportunity to develop a standardized set of questions that could be used to compare among study sites and interview groups. Few expectations about possible outcomes could be articulated in advance. The questionnaires were therefore designed to gather information from a diverse sample on a variety of topics. The questionnaires were drafted in English and translated to Thai by a native Thai speaker.
Data were analyzed comparing frequency distributions using chi square tests. To evaluate approval or disapproval for a response, observed frequencies were compared to expected frequencies, constructed using a random distribution of answers (20% in each category for 5-category answers, 50% in each category for yes or no answers). Observed distributions significantly different from random were considered positive if response bias was in the positive direction, and vice-versa. When overall results were significant, differences among response groups, provinces, or other categories were evaluated using the overall frequency distribution as expected and the group frequencies as observed. For all statistical tests, alpha was set at 0.05.
Three separate questionnaires were developed to gather information from farmers, head villagers, and other stakeholders.
A common set of questions regarding general social perception was asked of all groups. The farmer group included farmers who were currently engaged in shrimp culture, as well as those who had abandoned and converted land. Head villagers were local elected government officials who had responsibility for maintenance of certain local demographic records. In many cases, head villagers were also shrimp farmers. Other stakeholders were informants in ancillary industries, such as feed companies, or those unrelated to the shrimp industry. This group was included to assess perceptions of shrimp culture from those not directly involved, in order to determine if information gathered from farmers and head villagers differed from that of other members of the community.
Information was collected from a variety of sources on where there had once been shrimp farms, and signs of abandonment were used to generate farmer interviews. Head
villagers and other stakeholders were identified based on farmer information and knowledge of the local industry.
The spatial extent of various land uses in each study site was derived from two sources. When respondents could be located at places such as housing developments, active shrimp ponds, or sites of other aquaculture production, they were asked to provide estimates of the area of land used for each purpose. When no respondents could be found, the size of abandoned ponds was quantified using Geographic Positioning Satellite (GPS). During these site visits, GPS points were collected at each corner of an abandoned pond. These points were transcribed to a Geographic Information System (GIS) using ArcView, and the area of each abandoned pond was estimated.
Results
A total of 147 surveys were conducted, with the majority of interviews completed for farmers, and fewest for head villagers (Table 1). The distribution of interviews was fairly even across the three sites.
In all study sites, transitions from one form of aquaculture to another were made easily. Of the 109 farmers surveyed, 69% indicated they had ceased to farm shrimp at some time. In Chanthaburi, an additional six farmers had changed locations because of crop difficulties, bringing the total number of farmers who had temporarily converted or abandoned ponds to 74% of the sample. Slightly over one third of the farmers who had stopped at some point had returned to shrimp culture by the time of survey. These numbers demonstrated the prevalence of shifts in aquaculture production, as well as the difficulty in quantifying land dereliction at any point in time.
Multiple concepts of abandonment complicated the evaluation. From the researcher's perspective, abandonment occurred whenever a farmer ceased to seed shrimp into their pond for an indefinite period of time. Often, farmers did not consider their land abandoned or converted, even if they were not using it or were using it for alternate purposes. Therefore, it was difficult to assess which ponds were essentially fallow, temporarily converted, or truly abandoned.
For the purpose of this study, ponds were considered abandoned if they appeared to have been left for a significant amount of time with no activity or residents. Unused ponds were similar in that they were inactive, but resident farmers indicated they had indefinite plans for future uses. Fallow ponds were those inactive for a matter of weeks or months between crops, while truly converted ponds were being used for something other than shrimp culture, and the owner indicated no intention to return to shrimp. Land dereliction was used to refer generally to all types of abandonment
