Hello, I have some problems trying to write up the formula for lmer. I have 43 subjects ( random factor) which were seen twice ( Visit : repeated measure - fixed). on each visit the patient performed a graded effort exercise ( effort : repeated measures, ordered, fixed 4 levels). So Subject is a random factor, Effort is crossed with Visit and both are repeated measures, the efforts are ordered as it is a cummulative test. The model y ~ Visit * Effort + (Visit | Subject) + (Effort | Subject) seems appropriate but how can I get a linear-quadratic trend for effort? Also the Covar matrix for the repeated measures is 8 by 8 . It is not blocked beacuse there is the possibility of a learning effect between the visits. How can I specify this? Thanks for any help with this model. Heberto Ghezzo Ph.D. Meakins-Christie Labs Mc Gill University Montreal - Canada