Survey data for ex-post evaluation of the horticulture project in the Marshall Islands

Published: 27 January 2021| Version 1 | DOI: 10.17632/45fcrs8r8h.1
Contributor:
Yan-Tzong Cheng

Description

This Data provides supplementary information about how we explored aid effectiveness of the horticulture project in the Marshall Islands. The issue of sustainable development in small island developing countries has become increasingly important. We used a quasi-experimental design and selected the participants (using household as a unit) of the project to the experimental group and non-participants to the control group to evaluate the impact of the horticulture project. As for data collection in the field, we collected data through a structured questionnaire. We recruited one interviewer to conduct household surveys in the field for 40 days from April to June. A total of 96 valid questionnaires were collected, including 36 participants and 60 non-participants. The average household consumption of vegetables and fruits is 6.18 units (SD=7.84). Comparing project participants with non-participants, consumption of vegetables and fruits of the participants is 1.96 units higher than non-participants (P=0.062). Results from regression analysis indicate that participation in the project significantly affects consumption of vegetables and fruits (p< .05). Also, the revenue of selling vegetables and fruit accounts for 74.77 percent of participants' household income, becoming an important source of income for the households.

Files

Steps to reproduce

The horticulture project in the Marshall Islands was implemented during 2011–2014, and the three major components of the project: (1) provision of resources; (2) capacity building; and (3) vegetable and fruit promotion, achieved the outcome of more consumption of vegetables and fruits. In order to comply with the framework of quasi-experimental design, we selected and assigned participants (using household as a unit) of the projects to the experimental group and non-participants to the control group. We collected data through a structured questionnaire. We selected participants randomly from the participant list to conduct the household survey, and chose the non-participants according to their recommendations or from their neighbors who did not join the project. This survey mainly used households as a unit of analysis. We adopted snowball sampling within the network of project participants to select non-project participants. Non-project participants were recommended by project participants, and most of them were participants' neighbors. In addition, we had a local technical team to work with the local people; therefore most local people, including non-project participants, were willing to cooperate with us. With limited time and resources, we followed the rule of thumb and selected 1.5 to 2 times more non-participants than participants to ensure the intervention effect of the project can be explained. The respondents were conformed and did not include minors, indigenous peoples, pregnant women, persons with mental or physical disabilities, and people who are improperly coerced or unable to make decisions freely. We also designed several mechanisms such as strengthening the training of interviewers and locating the interviewees, to further mitigate selection bias. In addition, we collected data through a structured questionnaire designed by the TaiwanICDF and external consultants, using the intake of vegetables and fruits as the major measure variable to understand the intervention effectiveness. As for data collection in the field, we recruited an interviewer who was trained in basic interview skills and data entry to conduct household surveys in the field for 40 days from April to June. Then we interviewed the food preparer in the family to assess project effect as well as the family’s dietary pattern, with the food preparer usually providing more accurate informative answers. A total of 96 valid questionnaires were collected, including 36 participants and 60 non-participants. Finally, this data analyses include descriptive statistics and inferential statistics to verify the attribution of participating in the projects.