In all areas, scientific disciplines, in which it is possible to formulate conclusions from conducted research, confirm or refute scientific theses, etc. should be preceded by conducting specific research, analysis including also research, statistical tests. Of course, the type of research techniques and statistical study used is determined by the specificity of a specific field and scientific discipline, the specificity of the substantive content that is subject to research, and possible research techniques available in specific laboratories. On the other hand, the use of terms emphasizing the high level of significance of the problem and the use of such terms as "significant ..., significant ..., important ..., very important ..., particularly important ..., priority ..." it often includes a defined scope of subjective assessment. Therefore, these terms should be used to a moderate degree so that the specific substantive content contained in scientific texts is fully objective.

Greetings,

Dariusz Prokopowicz

Why not?

It should always be clear when "statistical significance" and when "practical/substantial/factual significance" (often termed "relevance") is meant. I am not a fan of writing/saying that some effect was significant, when it is actually meant that the data was statistically significant in a model assuming some particular effect (often a zero effect). That's just sloppy language, open doors for misinterpretations. But as long as author and reader know what is meant, there is no problem (however, I am quite sure that often neither author nor reader understand what is actually meant [*]).

As an extreme, take a case report: there is a single patient with some strange symptoms, and some treatment is just tried, because no standard option was available or helpful, it may turn out that after the treatment the symptoms got significantly better. Very significant for the patient, and a plausible reason to publish this case, so that others in a similar situation get a hint what might help. There is no statistics involved. To demonstrate that the cure is caused be the treatment, a study with a larger sample of such patients will have to be planned, in which one has to show that the observed effect is reasonably unlikely under a model that assumes at best no effect (what includes adverse effects). The data has thus to be statistically significant under that model to make that conclusion.

In practice, it would be more intersting (than just knowing the statistical significance under a null model) to have a reasonably precise estimate of the effect. Only this would enable us to judge if we would find that effect relevant or not. A very relevant effect may not give statistically significant data (if the variance is large and/or the sample size is small), and very statistically significant data may be obtained for irrelevant effects (if the variance is tiny and/or the sample size is huge).

---

[*]

• Gigerenzer, G., Krauss, S. & Vitouch, O. The null ritual: What you always wanted to know about null hypothesis testing but were afraid to ask. in Handbook on Quantitative Methods in the Social Sciences. Sage, Thousand Oaks, CA 389–406 (2004).
• Gigerenzer, G. Mindless statistics. The Journal of Socio-Economics 33, 587–606 (2004).
• Wasserstein, R. L. & Lazar, N. A. The ASA Statement on p-Values: Context, Process, and Purpose. The American Statistician 70, 129–133 (2016).
• Greenland, S. et al. Statistical tests, P-values, confidence intervals, and power: a guide to misinterpretations. European Journal of Epidemiology 31, (2016).
• Goodman, S. A dirty dozen: twelve p-value misconceptions. Semin. Hematol. 45, 135–140 (2008).

The word "significant" is an English word which may be used to describe an object which might not be necessarily "statistically significant" in some sense. Therefore there is nothing wrong in the use of the word in publications in a non-statistical sense.

Thank you Dr. Rosana Gonçalves Barros, Dr. Dariusz Prokopowicz, Dr. Jochen Wilhelm and Dr. Ette Etuk for your valuable contributions.

The problems is more for articles that do statistical tests, if they say "significant" without specifying whether they are using the English or Statistical meanings.

Thank you Dr. Daniel Wright for your response. I asked this question because I have read many articles that use the word significant a lot without testing, and for me the man cannot determine whether it is significant or not.

Use another word instead such as crucial, vital, important....

Thank you Sule Akkoyunlu for your answer.