Hi all, I am trying to fit simple 2 factor (factors A and B - 3 and 2 levels respectively) poisson regression model. Inititially a pure error model is fitted and significance tests are performed on each parameter. I wish to remove an individual parameter from the model - the interaction between the second level of factor A and factor B. However i only seem to be able to remove the AB (all interactions between A and B) term, which is no use as it also removes the interaction term between level 3 of Factor A and factor B. Can anyone help with this. Thanks Malcolm ---------------------- MJ Price, Social Medicine epmjp at bristol.ac.uk