Dear NBS experts,

Thank you for your great toolbox.

I would like to know how we can describe the NBS Method in an article (Method section of the paper)? Which parameter should we mention?
As you know, we have to summarize the method in one or two paragraphs.

Moreover, How can we report the statistical results derived from the NBS in the most evident manner(Result section of the paper)?
Some reviewers ask for more details about the statistical methods and results.

In advance, thank you for your kind attention.
Best regards.

Hi Omid, 

For the Methods section, I suggest reporting the test-statistic threshold that was chosen, and provide a very brief overview of the method. Check out some of the previous papers using the NBS for examples. A citation to the paper would also be very much appreciated. 

For the Results, I suggest reporting the size of any subnetwork, p-values and possibly also effect sizes. You could point the reviewer to the NBS paper for more details about the method.

It is difficult to provide more details without knowing the context and details of your study. 

Best wishes, 
Andrew

Originally posted by omid sefat:

Dear Andrew,

Thank you for your reply.


How can we compute the effect sizes ?

By computing effect size , we have one value for each significant component/subnetwork or for a significant edge?

Best regards,

An effect size can be define for each edge or for the whole subnetwork. 

Averaging the effect sizes across all the edges in a subnetwork is one way to estimate an effect size for the whole subnetwork. 

The NBS manual provides details about how to extract test statistics for individual edges. These test statistics can be converted to effect sizes. 

Originally posted by omid sefat:

Thank you dear Andrew,

Which Cohen's d formula do you suggest for the paired t-test ? 

Best regards,

d = t / sqrt(n)

where t is the t-statistic value and n is the number of subjects. 

You may need to multiply by a factor of sqrt(2) depending on the design. 
 
Originally posted by omid sefat: