Don't Believe Your Model Results

Mathematical modeling is a powerful way to study how fish populations behave in response to fishing, because fish populations are too remote to directly observe.  However, we should avoid an audacious belief in our model results.  We are much better served by recognizing the simplicities and assumptions inherent within models.  In addition to routine model validation and inspection of how well the model fits the data, we should consider what the data suggest on their own and compare the model results to independent information.  We should listen to critics of the model and people who do not believe the model results, so that we can scrutinize the model, testing as many of our assumptions as possible, to investigate whether or not the critics might be right.  Even after model validation, we should view the model as an imperfect tool and be open to other perspectives.  If a single, optimal model cannot be identified, information from multiple modeling approaches should be considered in scientific conclusions and management advice.  If we don’t defend models as though they’re gospel, we’re more open to applying alternative approaches and learning from our models.