Medical Science: Statistical Significance

Disclaimer: My knowledge on the subject is that of a layperson interested in science and health research.

"Significance" is a word that you'll come across while reading articles on medical research and studies, and it's easy to misinterpret what the word actually means. In statistics, "significance" does not imply importance, so saying that something is "significant" is ambiguous. The goal of this post is to explain the meaning, without using too much statistics jargon. A result of a study being "statistically significant" means that it is:

Likely that something occurred; probable that a detected difference is not by random chance; detectable.

Why?

Why would researchers use this type of statistical testing, if it doesn't give us answers to whether the results are important?

• It's easy to perform.
• It can measure properties that might be important in combination with other factors.
• It's much more likely that you get a "significant" result than if you only aim for measuring the effect size.
• It's a good first indicator of whether this line of research has potential. It provides a basis for further research.

Clinical significance

In everyday language, significance means "important", and there is a related concept in medical research: "clinical significance". The definition, according to Wikipedia:

the practical importance of a treatment effect — whether it has a ... genuine, ... noticeable effect on daily life

Further reading

Wikipedia has a couple of great articles on:

• "Statistical significance"
• "Clinical significance"

Corrections are appreciated!