Thinking with Computational Models to Guide Scientific Inquiry

Computational models make explanations precise enough to examine their quantitative consequences, or predictions, sometimes raising questions that are difficult to anticipate from verbal reasoning alone. To understand a prediction, we need to know what produced it, which assumptions it depends on, and what evidence could distinguish the explanation from alternatives. This tutorial explores how calculations and simulations can help us work through these questions. We begin with expectations about what a model will do and why, compare them with the results, and use what we learn to decide what to investigate next. A sequence of learning examples carries this reasoning into the design and analysis of a simulated experiment, including choices about what to measure and what the resulting data could establish. A decision-making example shows how two model settings can agree on accuracy but differ in decision times. Throughout, we explain why a particular comparison is informative and how its outcome guides the next question.