Good models: sound science, evidence and best-available data.
Good models are informed by sound science, empirical evidence and best-available data. They are also products of their designers’ worldviews, of the questions they presently think are relevant to their interests (often reflecting disciplinary commitments), and of the particular perspectives that happen to be on hand for addressing these questions. They always reflect contingent decisions about what to include and exclude. They can never provide complete accounts of the situations they represent. As George Box put it, ‘All models are wrong; some models are useful.’