Colin,
In order to create a contrast, you should use the following steps:
(1) Define your null hypothesis [e.g. insight=0];
(2) Make the null hypothesis equal to 0 [e.g. insight=0];
(3) Find the coefficients for each term [e.g. 1 for insight];
(4) Create a contrast of all zeros the length of the number of columns;
(5) Now fill in the coefficients from step 3.
Using this approach you will end up with a 1 over the covariate column
that shows positive correlation with insight. If you want the negative
correlation, then use a -1. If the first column is your covariate,
then the 1 or -1 should be above the first column. If the second
column is your covariate, then the 1 or -1 should be above the second
column.
Other notes:
(a) In the one-sample t-test all you have is a mean column. Thus, the
regression is a one-sample t-test with the covariate column added to
the model.
(b) The contrast [1 -1] is comparing the mean to the slope of the
covariate it. I do not know of any interpretation of the comparison of
the mean and slope of models.
Hope this helps.
Best Regards, Donald McLaren
=================
D.G. McLaren, Ph.D.
Postdoctoral Research Fellow, GRECC, Bedford VA
Research Fellow, Department of Neurology, Massachusetts General Hospital and
Harvard Medical School
Website: http://www.martinos.org/~mclaren
Office: (773) 406-2464
=====================
This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED
HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is
intended only for the use of the individual or entity named above. If the
reader of the e-mail is not the intended recipient or the employee or agent
responsible for delivering it to the intended recipient, you are hereby
notified that you are in possession of confidential and privileged
information. Any unauthorized use, disclosure, copying or the taking of any
action in reliance on the contents of this information is strictly
prohibited and may be unlawful. If you have received this e-mail
unintentionally, please immediately notify the sender via telephone at (773)
406-2464 or email.
On Mon, Apr 16, 2012 at 2:22 PM, Colin Hawco <[log in to unmask]> wrote:
> Hello,
>
> We wish to run a regression analysis on some data we have. This is our first time attempting such an analysis, and we want to make sure we do it correctly. We performed a contrast at the individual level (1st level analysis) contrasting 2 conditions (condition - control). We are now interested in examining how activity in our experimental condition (compared to the control condition) is related to a measure of insight we have taken from each participants.
>
> So, we performed a second level analysis with multiple regression, imputing our contrast files (con_0001.img) for each participant, and adding their insight score as the covariate (one per participant). This creates a model with the insight scores, and a mean column. I am not sure why this mean column is necessary, as it is not present in more simple analysis using one-sample t-tests.
>
> Our question is, what is the contrast for our results? As we have only 1 regressor (insight), i am presuming it should be [1, 0]. However, we tried using [1 -1] and it gives a much wider pattern of activity, but a pattern that still make a lot of sense. While my feeling is that the correct contrast is [1 0], I wanted to confirm that this was the case.
>
> So, for our regression analysis, should we be using the contrast [1, 0] or [1, -1] in our results?
>
> Thanks a lot for the help!
> Colin.
|