If, for example, I am running a GLM with Poisson distribution (it could be any distribution) and I have a Randomized Block Design (RBD) [Note: I don't wanna run a GLMM just to put the blocks as a random effect]. So my model would be VARIABLE ~ TREATMENTS + BLOCKS. Because of that, R estimates the coefficients for the treatments and the blocks. The problem is that in the intercept is adding the effect of the first treatment with the first block, and I am not interested in the block coefficients for the model.

How can I isolate the effects of the first treatment without the first block in the intercept?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.