Hi Jorge

Thank you for your answer our question about CM. I worked so far and the results are in line with glm_fit as you predicted.

If we would like to compare the groups (4th column in our qdec.table, with (responder = 1), (non-responder = 2) coded) the CM.C would be [0 0 1 0] ? We have unbalanced group sizes here 14 responder vs. 9 non-responder.

Thanks for your assistance

Best,

Vivian