Dear all, Lets say i have two groups (1 and 2) defined by an extern variable (a specific score calculated out of a questionaire). What gives me more explanatory power: A contrast between group 1 and 2 or a one sample t-test (2nd level analysis) using the score as covariate? Should there be differences in the contrasts and what does they mean? And an additional question: How can i define the grand mean for the covariate and should i use this option? Thank you again in advance! Best Marco