sdm-help-list > including mutliple contrasts
Showing 1-3 of 3 posts
Mar 25, 2017 01:03 PM | Jorge Witman
including mutliple contrasts
Dear expert,
I'm currently doing a meta-analysis with SDM and I decided not to include multiple contrasts of 1 study (say a>b and b > c) because this would lead to 1 study driving the results more than others (and non-independence of input studies). Am I correct with this assertion or does SDM allow for the inclusion of multiple contrasts of 1 study?
Moltes grĂ cies,
Jurriaan
I'm currently doing a meta-analysis with SDM and I decided not to include multiple contrasts of 1 study (say a>b and b > c) because this would lead to 1 study driving the results more than others (and non-independence of input studies). Am I correct with this assertion or does SDM allow for the inclusion of multiple contrasts of 1 study?
Moltes grĂ cies,
Jurriaan
Mar 30, 2017 08:03 AM | Joaquim Radua
RE: including mutliple contrasts
Dear Jurriaan
When a study includes multiple contrasts, one option is to only include the most relevant contrast as you did, but another option is to combine the multiple contrasts of the study in a single contrast (this can be done with the "Combine images" button). The first option may be adequate when one of the contrasts is clearly more representative than the others, whereas the second option may be more adequate to increase the power.
Moltes grĂ cies!
Joaquim
When a study includes multiple contrasts, one option is to only include the most relevant contrast as you did, but another option is to combine the multiple contrasts of the study in a single contrast (this can be done with the "Combine images" button). The first option may be adequate when one of the contrasts is clearly more representative than the others, whereas the second option may be more adequate to increase the power.
Moltes grĂ cies!
Joaquim
Apr 2, 2017 11:04 AM | Jorge Witman
RE: including mutliple contrasts
Thanks for the clear answer. Best, Jurriaan