Share your ideas for the JMP Scripting Unsession at Discovery Summit by September 17th. We hope to see you there!
Choose Language Hide Translation Bar
Highlighted
billi
Level IV

Effect summary- blue line significance

In effect summaries why blue line signifies significance is set at 0.01 not at 0.05? Is there an option to change it? Even if I select 'set alpha level' from red hotspot it did not change my results. Can anyone please explain this?

Thanks

3 REPLIES 3
Highlighted
P_Bartell
Level VI

Re: Effect summary- blue line significance

The blue line is drawn at 0.10, not 0.01 as you typed in your original discussion question. This line signifies the setting of the false discovery rate. In lay terms, the false discovery rate is a rate at which you can assess significance of effects that is especially useful when you are estimating lots of effects. It kind of guards against discovering falsely significant effects. The fdr method is contained in lots of JMP modeling methods. Here's a link to the JMP online documentation containing more details and a reference:

 

https://www.jmp.com/support/help/14/interactive-effect-summary.shtml#1534569

Highlighted
billi
Level IV

Re: Effect summary- blue line significance

Thank you Bartell for your response. I wrote 0.01 after reading the same page that you referenced. This line is from the same JMP documentation:- A value that exceeds 2 is significant at the 0.01 level (because -log10 (0.01)=2)).

Please correct me if I am wrong, bars (model effects) not crossing the blue lines are not significant. So the significance level is 0.01 here right?

Highlighted
P_Bartell
Level VI

Re: Effect summary- blue line significance

I will not say or ever use the two words 'statistical significance' together in talking about an effect as to having those properties. This is a matter of my personal and professional choice. There is lots of debate on those two words in the statistical inference construct...that I won't elaborate on here in this Forum. What I will talk about is the p- value and what it's interpretation. The documentation refers to the Logworth value of 2 corresponding to a p-value of 0.01. The blue line in the effect summary visualization is by default drawn at the p-value of 0.10. Whether or not an effect bar crosses that line depends on the Logworth value, and as such corresponding p value. It is left up to the individual interpreter to decide the question of 'statistical significance'.

Article Labels

    There are no labels assigned to this post.