A test for choice between hyperbolic sigmoidal models of crop yield response to weed density

The form of a crop yield-weed density response function influences control thresholds and risk management strategies. A choice of model test is introduced for evaluating whether yield response to weed density is sigmoidal or hyperbolic. A null hypothesis of hyperbolic yield response is evaluated in a Monte Carlo simulation of sigmoidal yield functions. The analysis considers two probability distributions of weed density and varying degrees of functional curvature, data variability, and sample size. The results suggest that, given empirical levels of functional curvature and data variability, sample sizes larger than the norm in weed science are required to reject the null hypothesis of hyperbolic response.