From Farmers to Breeders: Integrating video-based concept testing and data-driven text analysis to inform cassava breeding decisions

This study investigates how to integrate women's and men's cassava farming preferences into breeding strategies using Video-Based Product Concept Testing (VPCT). The data set contains preference rankings and qualitative responses from 398 cassava farmers in Colombia's Caribbean region (departments of Bolívar, Córdoba, and Sucre) who evaluated seven cassava product concepts co-developed with the International Center for Tropical Agriculture (CIAT) and the French Agricultural Research Center for International Development (CIRAD). Each product concept was presented via standardized one-minute videos to ensure clarity and minimize enumerator bias.
Methodology:Study design: This study employed a cross-sectional survey design using a randomized incomplete block approach adapted from the triadic comparison of technologies (tricot) model.

Sampling strategy: A total of 398 cassava farmers were selected using a random purposive sampling strategy (160 from Bolívar, 160 from Córdoba, and 78 from Sucre) to ensure participants had recent cassava-cultivation experience. The sample size was determined through power analysis using the formula N = f(α, β)² · (1/d)², with α = 0.05, 1−β = 0.95, and expected effect size d = 0.34, yielding a minimum required sample of approximately 380 observations to detect small-to-medium preference differences at 95% statistical power.