|
Inference-based model validation seeks to construct a statistical comparison of model
predictions against measurements of the target process. Previously, such validation
has used the hypothesis of no difference as the null hypothesis. This is unsatisfactory,
because tests are more likely to validate a model if they have low power. Here we introduce
tests of equivalence, which use the hypothesis of dissimilarity as the null. Thus, they flip
the burden of proof back onto the model. We demonstrate their application using an empirical
forest growth model and an extensive database of field measurements. Finally we provide some
simple power analyses to guide future model validation exercises.
|
|