The Methodology of Science

Background

At MStranslate, we are passionate about not only ‘translating’ science, but also helping improve the wider understanding of science and the foundations on which it is based.  Recently, we featured an article describing the ‘Principles of Science’ – as our second part in this series, we are going to discuss the “Methodology of Science”.

Methods in Science

Some of the features underpinning scientific methods are:

Reproducible

Methods must be clear and sufficiently detailed to allow others to do exactly what was done. It is not possible to conclude that the cake made by another does not taste the same as the original when a different recipe was followed.

You can’t dismiss the findings of Bake and Cakes (2015) who made a banana cake that tasted of bananas by making a similar cake but leaving the bananas out because Bake and Cakes (2015) neglected to mention this ingredient.

It is not possible to reproduce a scientific activity unless the steps, ingredients and processes are specified unambiguously to enable others to do exactly what was previously done.

Controlled

There are many factors in a scientific activity that must be controlled, only some of which are known. For these known factors, control can be exerted through design and sampling procedures, which impose necessary precision.

There are many more factors whose existence is unknown or whose role is unclear. This is where the use of a control group and the power of random allocation make fundamentally important contributions. Such factors can’t be excluded – you can’t exclude something of which you are unaware – but they can be ‘randomised out’ by being equally distributed between treatment and control conditions. This preserves the relative nature of the comparison but can affect the absolute magnitude of the treatment effect – can you work out why?

There is another aspect of the process that must be controlled – the eagerness or motivation of the researcher. In science, the desired end should never influence the means.

Reasoned

From the original idea, the method by which this idea is assessed must be reasoned and reasonable. Aspects of this may be mechanical, using precedent or accepted practice.

Other, and more important, aspects will involve a range of judgments – the ‘art’ of science.

For these aspects, there is no right answer, just ‘better’ or ‘worse’ answers. Blending these into a reasoned and defensible activity can be challenging, with shortcuts to be avoided at all times.

When in doubt, spell it out. And then, when completed, revise your approach based on this experience and constructive feedback from others.

Inflexible

The statement ‘When in doubt, spell it out’ was missing something.

And stick to it!

Once defined and detailed, it is essential to adhere to the prescribed method. It is preferable to finish what you set out to do and then make changes than it is to ‘tweak’ things along the way. Tweaks can be a source of bias in many ways, particularly if they are ‘hidden’ on the basis that they were harmless.

Adhere and be clear – if you’re a tweaker, you’re making your science weaker.