Ranking crop species using mixed treatment comparisons.

Area of Science:

• Agricultural Science
• Biostatistics
• Bioenergy Research

Background:

• Combining evidence from multiple studies is crucial for robust treatment comparisons.
• Mixed Treatment Comparison (MTC) offers a framework for synthesizing evidence across diverse trials.
• Ranking agricultural crop species requires analyzing complex yield data from numerous field experiments.

Purpose of the Study:

• To adapt and apply the Mixed Treatment Comparison (MTC) methodology for ranking agricultural crop species based on yield data.
• To develop and evaluate Bayesian MTC models for estimating yield ratios among bioenergy crops.
• To assess the consistency and reliability of MTC estimates compared to traditional statistical models.

Main Methods:

• A meta-analysis of yield data from 67 field studies involving 36 bioenergy crop species was conducted.
• Several Bayesian MTC models were developed using baseline treatment contrasts.
• Node-splitting was employed to assess estimate consistency, and results were compared with a two-way linear mixed model.

Main Results:

• The Bayesian MTC model incorporating study random effects and study-specific residual variances exhibited the lowest deviance information criterion (DIC).
• All tested MTC models that included study random effects produced comparable yield ratio estimates.
• The proposed Bayesian framework facilitated a detailed analysis of uncertainty in crop species ranking.

Conclusions:

• Bayesian Mixed Treatment Comparison models provide a robust framework for ranking agricultural crop species using yield data.
• The methodology allows for coherent judgments on crop performance by pooling direct and indirect evidence.
• The approach enhances the analysis of uncertainty in species ranking, aiding informed decision-making in bioenergy crop selection.