Convention for symbols indicating statistical significance?

Question

This question might probably also fit into Academia.SX. I'm writing my first paper in life sciences and I'm very often seeing asterisks as indicators for statistical significance (*).

I'm wondering if there's a convention on which symbol, and how many of them are to be used to indicate what the significance is, e.g. * for p < 0.05 and *** for p < 0.001?

score 11 · Accepted Answer · answered Nov 16 '12 at 19:25

11

Following is the convention:

ns      P > 0.05
*       P ≤ 0.05
**      P ≤ 0.01
***     P ≤ 0.001
****    P ≤ 0.0001

answered Nov 16 '12 at 19:25

abhinavkulkarni

876
1
8
15

1

Note that some used "+" specifically for results in-between 0.10 and 0.05. They are commonly termed "marginally significant". Keep in mind that such arbitrary thresholds are not really improving clarity of the results. It would be more fruitfull to elaborate on the effect size. – Denis Cousineau Aug 06 '22 at 19:23

score 8 · Answer 2 · answered Nov 01 '12 at 12:43

8

The convention is...

* yay, I can publish
** yay, I can publish and not get refuted
*** I have no idea what alpha means
**** my power is unfathomable
***** graphing the data and noting r = 0.98 wasn't good enough

Also, see the references in chi's answer.

answered Nov 01 '12 at 12:43

John

22,838
9
53
87

7

Because significance testing leads to so many problems let's convert to using skull and crossbones. – Frank Harrell Nov 01 '12 at 13:11

score 6 · Answer 3 · edited Apr 13 '17 at 12:44

6

This is correct, but please don't fall in the trap of the star system: The Earth Is Round, p < .05 :-)

edited Apr 13 '17 at 12:44

Community

1

answered Nov 01 '12 at 11:46

chl

53,725

Convention for symbols indicating statistical significance?

3 Answers3

Linked