The scientific method is essentially as follows — starting with a scientist or scientists making a large number of observations about some aspect of the world around them. For example, the effect of heating oil. To ascertain what happens when oil heats up, you’d expect that the process would have as little interference as possible, so you get a genuine idea of what the boiling point is, what happens to the oil and so on.

To do a test just once is problematic, because what if something was different the next time or the next time you tried? Therefore, you’d want to get as many examples of the experiment of boiling oil and document the data that you get, so you have a fair sample of what’s going on. Then it’s possible to create a theory, based on what evidence you have so far — and that can lead to a prediction as to what would happen if you ran the experiment of boiling oil again and again.

If in future tests, there was a discrepancy in the results, then the experiment could be checked and retested, and even allow for a modification in the theory if required. One of the benefits of there being established theories in science, we’re able to rely on certain findings and yet be open to changing them if there’s a new development or result.