help
81 Subscribers 
help > F test for a two-tailed t test
Showing 1-3 of 3 posts
Feb 15, 2018 12:02 PM | Alejandro Cosa
F test for a two-tailed t test
Dear NBS experts,
My experimental design consist in two experimental groups and we are interested in identifiyng that subnetwork significantly different between groups. My question is:
Despite the natural choice would be a ttest, is a F test suitable in the situation we dont hypothesize about the sign of the changes and increases in some subnetwork edges can be accompanied by a decrease in others?
Thanks very much in advance for your response.
Best whises,
Alejandro
My experimental design consist in two experimental groups and we are interested in identifiyng that subnetwork significantly different between groups. My question is:
Despite the natural choice would be a ttest, is a F test suitable in the situation we dont hypothesize about the sign of the changes and increases in some subnetwork edges can be accompanied by a decrease in others?
Thanks very much in advance for your response.
Best whises,
Alejandro
Feb 16, 2018 06:02 AM | Andrew Zalesky
RE: F test for a two-tailed t test
Hi Alejandro,
Yes - F-test is perfectly fine and your interpretation is correct. In this context, F-test is a two-sided alternative hypothesis, whereas t-test is one-sided.
Andrew
Originally posted by Alejandro Cosa:
Yes - F-test is perfectly fine and your interpretation is correct. In this context, F-test is a two-sided alternative hypothesis, whereas t-test is one-sided.
Andrew
Originally posted by Alejandro Cosa:
Dear NBS experts,
My experimental design consist in two experimental groups and we are interested in identifiyng that subnetwork significantly different between groups. My question is:
Despite the natural choice would be a ttest, is a F test suitable in the situation we dont hypothesize about the sign of the changes and increases in some subnetwork edges can be accompanied by a decrease in others?
Thanks very much in advance for your response.
Best whises,
Alejandro
My experimental design consist in two experimental groups and we are interested in identifiyng that subnetwork significantly different between groups. My question is:
Despite the natural choice would be a ttest, is a F test suitable in the situation we dont hypothesize about the sign of the changes and increases in some subnetwork edges can be accompanied by a decrease in others?
Thanks very much in advance for your response.
Best whises,
Alejandro
Feb 16, 2018 09:02 AM | Alejandro Cosa
RE: F test for a two-tailed t test
Dear Andrew,
Thanks a lot for your response.
Alejandro
Thanks a lot for your response.
Alejandro