3 - Scientific models, evidence and decisions

Developing and testing scientific models

A scientific model represents selected features of a system so that we can explain observations and make testable predictions.

A particle diagram is a representational model: circles stand for particles and their arrangement represents a substance. A three-dimensional molecular model is a spatial model that makes shape easier to visualise. A verbal account of collisions is a descriptive model. An equation linking measured quantities is a mathematical model. A computational model uses a computer to apply rules and calculate many consequences, for example the effects of changing several atmospheric conditions. Each type makes some features easier to study while simplifying others.

The circles in a particle diagram do not show the true size, motion or forces of every particle. A model's usefulness depends on whether its features match the question being asked. An arrangement diagram can explain why a solid keeps its shape, but cannot by itself predict a precise melting temperature. To test that prediction we need further information about the substance and measurements made under stated conditions.

Scientific explanations develop through a cycle. Observations suggest an explanation; the explanation predicts something observable; a planned investigation tests that prediction; the results are compared with it. If several well-controlled investigations disagree with an important prediction, the explanation may need changing. New instruments can provide evidence that was previously unavailable. The atomic model changed when scattering observations showed that a model with positive charge spread throughout an atom could not explain the large deflections of a few particles. A concentrated positive nucleus explained those observations more successfully.

A successful prediction supports a model within the conditions tested. It does not establish that the model describes every situation. Two models can make the same prediction in one situation; choose another observation that would distinguish them. Check the method and uncertainty before treating one surprising result as a reason to discard an established explanation.

Evidence, technology and ethical choices

Science can investigate how a technology works and its consequences, while decisions about its use also involve priorities and values.

A water-treatment technology might remove harmful microbes while using electricity and producing waste. Chemical evidence can establish the removal achieved under test conditions, the energy used and the substances left behind. Measurements may be uncertain, the study may cover only one type of water, and long-term effects may take time to investigate. A short successful trial leaves those questions open. Explaining what is unknown helps a decision-maker choose further evidence to collect.

An evaluation compares personal, social, economic and environmental consequences. Personal consequences include effects on a user's health or convenience. Social consequences include access for different communities. Economic consequences include purchase and running costs. Environmental consequences include resource use, emissions and waste. A benefit in one category can come with a cost in another. State the comparison and its conditions: a more expensive filter could be a sensible choice where its improved treatment prevents a serious problem, but the same expense might prevent other households gaining any treatment.

An ethical issue concerns what ought to be done, including rights, fairness and responsibility. For example, if a technology's waste is released near people who receive none of its benefits, a question of fairness arises. Scientific measurements can establish the likely exposure and harm; they cannot alone decide how that burden ought to be shared. A reasoned ethical argument identifies the affected groups and the principle being used, then relates it to the evidence.

When information is supplied, use its quantities and limitations. An advertising claim about “less waste” needs a comparison per equivalent service, such as per litre of water treated. The strongest decision explains why selected benefits outweigh selected disadvantages for the stated purpose. A valid alternative decision can be justified if it gives different reasonable weight to the consequences.

Hazard, risk and the perception of risk

A risk judgement considers both the chance of harm occurring and the seriousness of its consequences.

A hazard is something with the potential to cause harm, such as a corrosive solution. Risk depends on exposure, the likelihood of an incident and the consequences. The same solution presents different risks when a small amount is handled with suitable eye protection and when it is splashed carelessly. Reducing exposure or the chance of a spill reduces risk even though the substance still has hazardous properties.

Estimating the frequency of a leak

Illustrative trials: one device leaks in 12 of 1200 uses. Divide incidents by uses, then convert the proportion to a percentage.