BiologybioRxiv
Heuristic editor, no API keyVerdict: NotablePlanted origin of shade trees: a robust determinant of cocoa yield among smallholder farmers in Cote d'Ivoire
CONTEXT: While the agroeconomic literature on the determinants of cocoa yield among West African smallholders is abundant, it remains dominated by descriptive studies or work conducted at an aggregated regional scale…
Key numbers
- 5-fold cross-validation
VerdictWorth a reader's time today.
Abstract
CONTEXT: While the agroeconomic literature on the determinants of cocoa yield among West African smallholders is abundant, it remains dominated by descriptive studies or work conducted at an aggregated regional scale, leaving open the question of the role of shade-tree management practices at the scale of the individual plantation. OBJECTIVE: This study tests whether shade-tree management practices and the farmer's socio-demographic profile explain variation in cocoa yield, using a nationwide sample of 409 plantations covering, for the first time, the three major Ivorian production zones (loops) (98, 151 and 160 plantations for loops 1, 2 and 3, respectively). METHODS: A univariate screening of 27 variables, a mixed model with a random intercept by village, a production-function specification, LASSO variable selection, and a random forest were applied, all evaluated by 5-fold cross-validation. RESULTS AND CONCLUSIONS: The five approaches converge on a robust result: the deliberately planted origin of shade trees (as opposed to a residual or spontaneous origin) is the strongest and most stable determinant of yield, with a mean gap of 332 versus 183 kg/ha/year. This effect withstands four successive robustness checks: it remains significant after simultaneous adjustment for age, plantation size, production zone, technical extension, and farmer education; it is not driven by a handful of extreme plantations; it holds within each of the three production zones taken separately rather than in only one of them; and, taken in isolation, it retains a positive out-of-sample predictive power (cross-validated R2=8776; 0.05). A second group of robust determinants of more modest magnitude emerges for agricultural technical extension (ANADER/CNRA/SATMACI) and farmer education level; shade-tree alignment shows a signal in the same direction, consistent with recent independent work in Cote d'Ivoire, but becomes statistically marginal once adjusted for these other factors. The full multivariate model reaches a cross-validated R2 of around 0.07-0.08. SIGNIFICANCE: This signal, undetectable in an analysis restricted to loop 1 alone (n=98) for lack of statistical power, confirms the value of nationwide sampling for detecting modest but real agronomic determinants, and argues for integrating shade-tree origin into agroforestry extension programmes.
The editor's rubric
| Dimension | Level | Weight | What that level means |
|---|---|---|---|
| Leverage | ███░░ 3 | 24% | A method or resource many groups across the field will adopt within a year. |
| Magnitude | ████░ 4 | 16% | A qualitative jump: a capability or regime that did not exist before. |
| Evidence | ███░░ 3 | 20% | Solid: multiple benchmarks or cohorts, ablations, fair baselines, released code or data. |
| Novelty | ███░░ 3 | 20% | A genuinely new approach to an open problem. |
| Trajectory | ███░░ 3 | 10% | A clear path to scale. |
| Stakes | ██░░░ 2 | 10% | Benefits a professional community (practitioners, clinicians, engineers). |
Editor’s rationale
Heuristic triage from title and abstract text only, not a reading of the paper. Cues found: breadth (many tasks); gains (x-fold, versus baseline); firsts (for the first time, first); novelty (alternative to status quo); verification (experimental validation, independent replication); scale (scalable, improves with scale).
How the score was computed
- Merit
- 6.1 / 10
- Adjusted merit
- 5.1 / 10
- Attention
- 0%
- Freshness
- 78%
- Citations0 (reference 20, via openalex, Sep 29, 2026, 23:37 UTC)