Hi Susanne,

It seems the comparison between conditions is being done at the higher level (intermediate, 2nd level), is this the case? If yes, then anything that is specific to that condition is already accounted for by the EVs that encode that condition, and there is no need to add the further EVs for the conditions' response time. Even though these aren't explicitly included, they are already captured by the model.

If, however, you have a different setup, please attach the "design.png" for both the lower level and for the higher level so we can have a look.

All the best,

Anderson


On 2 October 2015 at 07:49, Susanne Sommer <[log in to unmask]> wrote:
Hello everyone,
I am having troubles regarding an analysis I want to do. I have a relative simple experiment with 4 conditions. Now, reaction times are different depending on conditions. I tried to account for that through modeling EV duration from onset 'til time of reaction. Now I would like to add the average response time per condition as a regressor in the model but I am unsure as to how to accomplish that. I tried to add 4 regressors on higher level that included the demeaned average response time  per condition per person (demeaning done per person by subtracting mean from mean of condition) but I got the error message, that one or more EVs are close to a linear combination of the others. Now I am unsure as to how to incorporate the mean reaction time per condition.
Thanks for your help,
Susanne Sommer