Developing and Reviewing Scientific Theories

From observations to explanations

Scientific observations tell us what happens; explanations suggest why it happens. An explanation does not emerge automatically from a collection of results. Scientists use creative thinking to propose ideas, then work out how to test them.

A scientific theory is an evidence-supported explanation that connects observations and makes testable predictions. In science, ‘theory’ does not simply mean a guess. A well-established theory has survived many tests, but it can still be modified if convincing new evidence becomes available.

Scientists might begin with a question, propose a hypothesis and predict what they should observe if it is correct. They then investigate and compare the results with the prediction. Agreement supports the explanation; disagreement prompts them to check the investigation and consider whether the explanation needs changing. One unexpected result is not automatically enough to overturn a theory: errors and alternative explanations must also be considered.

How methods and theories develop together

Scientific methods develop as scientists find new ways to investigate questions. Improved equipment can reveal things that earlier methods could not detect, allowing explanations to be tested more thoroughly.

The development of ideas about atoms shows this connection. An atom is a tiny particle of an element. In 1803, John Dalton presented an atomic theory in which matter consisted of tiny, indivisible atoms. Different elements had different atoms, and atoms combined to form new substances. This provided a useful explanation of matter, but later experiments showed that atoms were not indivisible.

In 1897, J. J. Thomson investigated beams in a cathode-ray tube. Their deflection towards a positive plate provided evidence for negatively charged particles: electrons. Finding particles smaller than an atom meant that Dalton’s description needed changing.

Thomson proposed the plum pudding model: negative electrons embedded in material with positive charge spread throughout it. This accounted for the presence of electrons while explaining how an atom could have no overall charge.

Unexpected scattering changes the explanation

In the gold-foil experiments carried out by Geiger and Marsden under Rutherford’s direction in 1909, a beam of positively charged particles was directed at thin gold foil. This was a different investigative method: rather than studying a beam of electrons, scientists used particles to probe the structure of atoms.

With positive charge spread throughout an atom, the plum pudding model did not predict the very large deflections that were observed. Most particles passed through the foil, but some changed direction and a very small number were deflected back.

These observations supported a new explanation:

• Most particles passed through: atoms must be mostly empty space.
• Some particles were strongly deflected: positive charge must be concentrated rather than spread throughout the atom.
• A few particles were deflected back: this concentrated region must be small and contain most of the atom’s mass.

The resulting nuclear model placed a small, positively charged nucleus at the centre of the atom, with electrons outside it. The important development was not simply that a scientist suggested a different picture: the new description explained evidence that the earlier description could not.

In 1913, Niels Bohr further developed the nuclear model by proposing that electrons occupy shells at set distances from the nucleus. His calculations agreed with experimental results. Scientific explanations can therefore develop in stages: useful parts of an earlier explanation may be retained while other parts are changed.

How peer review checks scientific claims

Before a new claim is widely accepted, other scientists examine the work behind it. Researchers write a scientific paper describing their question, methods, results and conclusions and submit it for publication.

In peer review, other scientists with relevant expertise assess the paper. They consider whether the methods are suitable, whether the evidence is convincing and whether the conclusions are justified by the results. For example, a reviewer might question whether a fair comparison was made or whether the authors have claimed more than their data support.

Reviewers may recommend publication, request changes or recommend rejection. Authors may need to explain their choices more clearly, correct errors or provide further evidence. This checking helps improve the quality and trustworthiness of published research.

Acceptance remains open to further testing

Peer review is not the same as repeating an experiment, and it cannot guarantee that a claim is correct. Reviewers may miss mistakes, bias or manipulated data. Other scientists can carry out independent investigations to check whether the findings hold up.

Confidence grows when different investigations provide consistent, strong evidence. Nevertheless, an accepted theory remains open to testing. If reliable new evidence conflicts with it, scientists may modify or replace it. This willingness to change explanations in response to evidence is a strength of science, not a sign that all scientific knowledge is equally uncertain.

Developing explanations

• Observations are evidence; theories explain them. Proposing explanations involves creative thinking.
• Question → hypothesis → prediction → investigation → comparison with evidence.
• New methods and equipment enable new tests. Strong conflicting evidence can lead to a theory being modified or replaced.

Atomic example

• Dalton, 1803: atoms described as indivisible.
• Thomson, 1897: discovery of negative electrons showed that atoms contain smaller particles; plum pudding model proposed.