Hi Jennifer,

Yes, it is rather cumbersome but possible to define a paired t-test in that context, you just need to define two between-subject factors identifying the two conditions, and then explicitly define all subject-effects (one per subject) identifying which participants in your project correspond to data from the same person. See the following post for details: https://www.nitrc.org/forum/message.php?...

note: in that post example code the data is sorted differently than yours (first all subjects pre condition, then followed by all subjects in post condition), in your case you would just need to define instead:

group_pre=1:2:100;     % i.e. 1, 3, 5, etc.
group_post=2:2:100;   % i.e. 2, 4, 6, etc.

changing the number 100 above to twice your number of real participants.

Hope this helps

Alfonso

Originally posted by Jennifer Siegel :

Happy New Year’s Conn Users,

We have set up a pre and post-analysis in Conn.  We’ve set this up so that each pre and post-session is a separate participant (see example below).  I understand it’s better to do this with each participant having pre and post and indicating these differences in the conditions tab.  Given that we have already set this up, is it possible to run a within-subjects analysis at this point through the gui? 

Participant 1 = subject 001 – pre

Participant 2 = subject 001 -post

Participant 3 = subject 002 – pre

Participant 4 = subject 002 – post

Thanks for any help you can provide,

Jen