Our activities often encourage students to rule out poor models, even before they can decide on a best model.
A very important (and under-appreciated) scientific activity is ruling out models that are strongly contradicted by data. Even when it is not possible to settle on a best model, it is often possible to rule out poor models (because of lack of fit with data, with established theories, etc.).
Therefore, some PRACCIS activities focus not on choosing the best model but rather ruling out models. For example, in our natural selection unit, students are asked to consider whether a series of models fits the evidence.
In each case, one core feature of the model is strongly contradicted by the evidence, so the students learn that they can rule these models out, even if they are not yet ready to adopt a correct model. In the example on this page, the model fails to explain evidence that horn size in mountain sheep is an inherited trait.