hi - think this is a pretty basic question so apologies for that We are interested in modelling a learning effect in a simple desigm Each participant has 2 runs, 5 blocks each For each block we modelled EVs for predictive and unpredictive cues, leading to 10 EVs For the first-level analyses, we computed 5 copes for the difference between predictive > unpredictive cues We are interested in the learning effect, namely how the difference between predictive and unpredictive cues changes across the 10 blocks We then submit, for each subject, the 10 copes from the 2 first level-analyses into a Fixed effects analyses specifying the design as 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 and (for example) a linear contrast -4.5 -3.5 -2.5 -1.5 -0.5 0.5 1.5 2.5 3.5 4.5 I guess this model look fine? or would it be more appropriate to do a Fixed effects with one EV for the relevant cope and another covariate to model the learning effect, such as 1 -4.5 1 -3.5 1 -2.5 1 -1.5 1 -0.5 1 0.5 1 1.5 1 2.5 1 3.5 1 4.5 and then assess the covariate effect for each subject and bring that to the 3rd level which model would be more appropiate in this case? guess the 2 models are pretty much the same..what is the difference if any or is there a better way to model for this experiment? suggestions welcome and appreciated! thanks so much David