Only 0.7% of Agricultural Research Reports Link Trust to Scaling

Only two reports out of 273 — a mere 0.7% — document social capital and scaling together in a meaningful way. That's the headline number from a new study in Frontiers in Sustainable Food Systems, which examined how agricultural research reports handle the messy human side of spreading innovations to farmers.

The finding matters because getting innovations to scale — reaching more farmers, shaping policy, changing practices — is one of the hardest problems in agricultural research for development. And the study's authors argue it depends heavily on trust, partnerships and collaboration, not just better seeds or smarter techniques.

Who did the study, and how?

Deissy Martínez-Barón, a researcher at the International Center of Tropical Agriculture (CIAT, now part of the Alliance of Bioversity International and CIAT) and the University of Copenhagen, worked with Anne Gravsholt Busck of the University of Copenhagen and Steven Dean Prager of the Gates Foundation. Together they analyzed 273 results reports covering 261 projects across 12 CGIAR research programs, all reported between 2018 and 2021 under the former CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS).

The team used a text-mining method called latent Dirichlet allocation, or LDA — a statistical model that detects recurring themes in large collections of documents. The model surfaced 17 topics, which the researchers grouped into four areas: agricultural systems, climate change adaptation, climate change mitigation, and enabling environments.

What did the numbers show?

The pattern is striking, though the authors are careful about what it means:

• Scaling itself is barely visible as a topic. The topic most directly tied to scaling showed a significant presence in only 1.8% of the 273 reports. That doesn't mean projects skipped scaling activities — it means the documentation rarely treated scaling as a process worth explaining.
• Social capital shows up, but under another name. Topics tied to enabling environments — collaborative research, institutional partnerships, stakeholder participation — appeared significantly in 45.4% of reports. These reflect social capital in spirit, though the authors concede they don't directly measure trust between actors.
• The two almost never meet on the page. Those two reports, 0.7% of the total, were the only ones showing a significant simultaneous presence of both social capital and scaling topics.
• Social dimensions vary by field. Adaptation research emphasized capacity strengthening and climate information services, while mitigation work leaned toward technical practices and policy instruments.

The study frames scaling in three pathways: scaling out (more people and places), scaling up (policies and institutions), and scaling deep (transformed relationships, practices and norms). Reports mostly documented the first two, and almost always as achieved outcomes rather than as processes. Scaling deep got the least attention of all.