 |
 |
Hi Ting, Yes, it's possible. In the design shown in the other thread, i.e., the one in which there are EVs for the subject-specific slopes, remove the first 2 EVs (that coded group) and replace for: - EV1: intercept - EV2: BDI The contrasts are then [0 1 0 0 ... 0] and [0 -1 0 0 ... 0]. All the best, Anderson On 11 June 2016 at 16:54, Li, Kim <[log in to unmask]> wrote: > But I want to see how BDI scores across subjects will affect their > reaction to the drug, so that's actually looking for a between subject > effect of the covariate. Is that possible to do in paired t test? > > Ting > > Sent from my iPhone > > On Jun 11, 2016, at 3:07 AM, Anderson M. Winkler <[log in to unmask] > <[log in to unmask]>> wrote: > > Hi Tin Li, > > In this case, with the same value for both visits, and considering that > the paired t-test is to test within-subject effects, there is no need to > include that covariate, as it's already taken care of by the > subject-specific EVs. So, just drop this variable, say, BDI, and yet BDI > (and all other things that didn't change between visits) will be taken into > account regardless. > > All the best, > > Anderson > > > On 10 June 2016 at 21:41, Ting Li <[log in to unmask]> wrote: > >> Hi All, >> >> In my study, subjects come twice, one time they receive drug and another >> they receive placebo. We are interested in the difference between drug and >> placebo, so we did a paired t test. Now, we have some behavioral measures >> we want to put in the model to see how they modulate the brain reaction. >> If, for example, someone with higher BDI will have stronger reaction in >> some parts of the brain for drug-placebo contrast. >> I have set up the model as shown in the attachment, with paired t test + >> one more EV. >> Because the same subject comes twice but each person only has one >> behavioral result, so the second half of the covariate is just the original >> covariate number * -1. (Is this done correctly? or should I put zeros?) >> Now, what should I do for the contrast to see the covariates' effect? Is >> c[100001] correct? >> >> 1 1 0 0 0 2 >> 1 0 1 0 0 4 >> 1 0 0 1 0 -5 >> 1 0 0 0 1 -1 >> -1 1 0 0 0 -2 >> -1 0 1 0 0 -4 >> -1 0 0 1 0 5 >> -1 0 0 0 1 1 >> >> >> > > ------------------------------ > > This e-mail message (including any attachments) is for the sole use of > the intended recipient(s) and may contain confidential and privileged > information. If the reader of this message is not the intended > recipient, you are hereby notified that any dissemination, distribution > or copying of this message (including any attachments) is strictly > prohibited. > > If you have received this message in error, please contact > the sender by reply e-mail message and destroy all copies of the > original message (including attachments). >