(P26-023-23) Getting Good Data: Insights From an Agricultural Production Survey

Objectives: Evaluating the effectiveness of any nutrition-sensitive agricultural (ag) intervention requires collecting in-depth and complex ag data. These data include aspects of land availability over time, normal farming practices, and yield for all types of crops grown on the farm, all framed in relation to how food is used by a household and the quality of dietary patterns that emerge. We identify the best practices and approaches to better integrate ag surveys into nutrition assessments.

Methods: As part of a cluster randomized intervention (One Nutrition in Complex Environments), we implemented two social behavior change (SBC) approaches of which one included an intensive ag intervention. Evaluation of the SBC approaches thus required collecting ag survey data. This was via in-person interviews at baseline and endline. Records were maintained of the survey design process, data collection tool programming, enumerator training, and on the ground support during survey implementation and synthesized. Challenges in collecting data from innumerate respondents, recall and social desirability bias, and participant attrition were identified and mitigated.

Results: Ag data collection is complex and time consuming and if not managed can yield unusable data. Firstly, data collection must follow the seasonal calendar which varies by the crop being assessed. Timing of ag surveys may vary from nutrition assessment time points and should be accounted for in planning and implementation. Ag surveys are extremely context specific and need to be validated and/or contextualized to the agro-ecology.

Collection of reliable data requires setting an electronic data collection portal that can handle the in-depth nature of inquiry. Training is critical as data collectors may not be versed in ag economics. There is also need for detailed training materials and rigorous training and re-training. During data collection, daily monitoring of incoming data is required as well as observing interviews and examining potential errors in the data.

Conclusions: Successful collection of reliable data rests on detailed planning before, and ongoing monitoring during, an intervention trial. We identified key learnings regarding major steps in design, planning, training, and data collection.

Funding Sources: U.S. Agency for International Development (USAID) and the Feed the Future Initiative