help
help > post-hoc group contrast analysis
Feb 1, 2023 02:02 AM | Lenore Tahara-Eckl - University of Auckland
post-hoc group contrast analysis
Hi Everyone,
I wanted to verify the post-hoc analysis process that I am running. For context, I have 5 groups and have done a between-group difference contrast ([-1 1 0 0 0; 0 -1 1 0 0; 0 0 -1 1 0; 0 0 0 -1 1]). After running this, I got two significant clusters with Functional Network Connectivity.
From here, I want to know where the differences are between the groups. I looked at the 'plot effects' graph in the Results Explorer window, which seems to indicate that group 1 is different from group 5 (i.e., group 1 > group 5). Following previous forum posts (https://www.nitrc.org/forum/message.php?... and https://www.nitrc.org/forum/message.php?...), I exported a mask to save the two significant clusters, and then I selected group 1 and 5 to conduct between-subjects contrast on the CONN gui. Then I re-ran the 2nd-level analysis and selected 'other clusters of interest' to input the saved mask file. From there, I get the statistics (e.g., t-value, p-value) between the two groups that I am comparing.
Is this the correct process for conducting the post-hoc group comparison analysis? And if I wanted to compare other groups to one another, then I would repeat this process, but with selecting those specified groups? Thank you.
I wanted to verify the post-hoc analysis process that I am running. For context, I have 5 groups and have done a between-group difference contrast ([-1 1 0 0 0; 0 -1 1 0 0; 0 0 -1 1 0; 0 0 0 -1 1]). After running this, I got two significant clusters with Functional Network Connectivity.
From here, I want to know where the differences are between the groups. I looked at the 'plot effects' graph in the Results Explorer window, which seems to indicate that group 1 is different from group 5 (i.e., group 1 > group 5). Following previous forum posts (https://www.nitrc.org/forum/message.php?... and https://www.nitrc.org/forum/message.php?...), I exported a mask to save the two significant clusters, and then I selected group 1 and 5 to conduct between-subjects contrast on the CONN gui. Then I re-ran the 2nd-level analysis and selected 'other clusters of interest' to input the saved mask file. From there, I get the statistics (e.g., t-value, p-value) between the two groups that I am comparing.
Is this the correct process for conducting the post-hoc group comparison analysis? And if I wanted to compare other groups to one another, then I would repeat this process, but with selecting those specified groups? Thank you.