help
81 Subscribers 
help > Equivalence test
Showing 1-2 of 2 posts
Jun 28, 2020 11:06 PM | thomas88
Equivalence test
Dear Andrew,
I am curious if there is any way within NBS to perform an equivalence test (i.e., to check for equivalence between two groups and thus provide support for the absence of a meaningful effect)? For example, say I have two groups which do not show significant differences in connectivity at baseline (p > .05), but then show differences after an intervention, is there any way that I can check that the groups were equivalent at baseline? I know this is something that can be done relatively easily if I were using a traditional t-test, however I'm not sure if it is possible to check given the added complexities of the NBS approach? Thanks.
I am curious if there is any way within NBS to perform an equivalence test (i.e., to check for equivalence between two groups and thus provide support for the absence of a meaningful effect)? For example, say I have two groups which do not show significant differences in connectivity at baseline (p > .05), but then show differences after an intervention, is there any way that I can check that the groups were equivalent at baseline? I know this is something that can be done relatively easily if I were using a traditional t-test, however I'm not sure if it is possible to check given the added complexities of the NBS approach? Thanks.
Jun 29, 2020 09:06 AM | Andrew Zalesky
RE: Equivalence test
Hi Thomas,
Equivalence tests are a great idea and I have enjoyed reading the recent work by Lakens on this topic.
I think that performing an equivalence test under the NBS framework would require tweaking of the code and would not be possible using the GUI itself. However, I have not given this a lot of thought. The definition of equivalence for networks is also a bit more tricky than for a t-test.
Future research that might be taken up by others!
Best wishes,
Andrew
Originally posted by thomas88:
Equivalence tests are a great idea and I have enjoyed reading the recent work by Lakens on this topic.
I think that performing an equivalence test under the NBS framework would require tweaking of the code and would not be possible using the GUI itself. However, I have not given this a lot of thought. The definition of equivalence for networks is also a bit more tricky than for a t-test.
Future research that might be taken up by others!
Best wishes,
Andrew
Originally posted by thomas88:
Dear Andrew,
I am curious if there is any way within NBS to perform an equivalence test (i.e., to check for equivalence between two groups and thus provide support for the absence of a meaningful effect)? For example, say I have two groups which do not show significant differences in connectivity at baseline (p > .05), but then show differences after an intervention, is there any way that I can check that the groups were equivalent at baseline? I know this is something that can be done relatively easily if I were using a traditional t-test, however I'm not sure if it is possible to check given the added complexities of the NBS approach? Thanks.
I am curious if there is any way within NBS to perform an equivalence test (i.e., to check for equivalence between two groups and thus provide support for the absence of a meaningful effect)? For example, say I have two groups which do not show significant differences in connectivity at baseline (p > .05), but then show differences after an intervention, is there any way that I can check that the groups were equivalent at baseline? I know this is something that can be done relatively easily if I were using a traditional t-test, however I'm not sure if it is possible to check given the added complexities of the NBS approach? Thanks.