As practitioners, we are not aware of the models that construct the datasets we study. We often rely on intuition to determine which model to choose. This white-paper takes a look at a penitential pitfall: What happens when we use a different model from the actual model generation process?