Cognitive models play an important role in increasing our understanding of cognitive systems, and will only become more important as computing power grows and models themselves become increasingly powerful. Despite this, the value of cognitive models can at times be viewed with some uncertainty and even skepticism. To address current issues with and critiques of cognitive model use, and to support cognitive scientists seeking to create better cognitive models and use these models more effectively, we provide a detailed picture of how such models fit into the scientific research process and, in doing so, offer a more concrete explanation of the appropriate functionality of models within cognitive science. As a further step towards clarifying the role of models in this area we have also identified two categories of models- explanatory and predictive- that can provide criteria for evaluating models and model use. By choosing appropriate model types and following these criteria, cognitive modellers will create more powerful models and use them with greater success.