5 Models and Representation

What scientific models represent

A scientific model represents a selected object, process, or system—its target—to investigate or communicate something about it. Models can take the form of physical replicas, diagrams, mathematical equations, conceptual descriptions, or computer simulations.

A model preserves or highlights features relevant to a question rather than copying every detail. The features that matter depend on what the model is being used to investigate. For example, a map can show roads and distances while leaving out trees, buildings, and elevation. Different models of the same system can therefore be useful for different purposes.

How models simplify

Models simplify through abstraction, which leaves out details, and idealization, which deliberately assumes simplified conditions that may not hold exactly in reality. These choices can make systems easier to analyze, explain, or calculate, but they also limit what a model can reliably tell us.

For example, a basic model of a planet’s orbit might treat the planet and star as point masses and ignore the gravitational effects of other bodies. It can help explain or predict the orbit when those omitted effects are small enough for the task. It is not a complete description of the solar system. If the question concerns a small orbital deviation, the omitted effects may become important and the model may need refinement.

A model is not simply good or bad in isolation. Its usefulness depends on what it represents, the purpose for which it is used, and the conditions under which it is applied. A model that works well for one scale or question may be inadequate for another.

How models support investigation

Scientists use models to reason about systems that may be too large, small, complex, distant, or slow to study directly. A model can help them:

• Describe structure: identify a system’s parts, boundaries, and connections.

• Explain behavior: show how interactions among parts could produce an observed pattern.

• Make predictions: derive expected outcomes under specified conditions.

• Explore possibilities: vary an input or assumption and examine what changes.

• Guide investigation: identify observations or measurements that could distinguish competing accounts.

For instance, a computer model of a spreading disease can represent interactions among people and explore how changing contact patterns might affect transmission. Its output is not a direct observation of the future; it depends on the model’s assumptions, inputs, and representation of the population. Comparing predictions with evidence helps scientists assess where the model works and where it needs revision.

Models range from simple sketches to detailed simulations. Useful system models make clear which parts and interactions they include.

Evaluating models

Evaluate a model in relation to its purpose and the available evidence. Questions to consider include:

1. Is the target and scope clear? Identify what system or aspect the model represents and what is outside its boundary.

2. Are the assumptions appropriate? Consider whether its simplifications preserve the features relevant to the question.

3. Does it agree with observations? Examine how well its results match available measurements.

4. Where does it work? Identify the conditions, scales, or situations within its useful range.

5. How uncertain are its results? Consider which assumptions or data could substantially change the conclusions.

A model does not prove itself true merely because it is useful. Evidence supports or challenges particular uses of a model. As questions and evidence change, scientists may compare, revise, or replace models.

From models to the world

Representation involves interpretation: scientists use a model as a stand-in for its target. They can infer from the model to the world only where the relevant connections are justified.

Scientific models help people describe, explain, predict, and investigate systems. Their simplifications make inquiry possible while setting limits on their use. Assessing a model requires attention to its purpose, assumptions, evidence, scope, and uncertainty. A model is a tool for learning about the world, not the world itself.