I'm fairly new to GLM and would like some advice on the lognormal model with glm() in R.

I have a bunch of zeros in my data and I realised that there are models out there like the zero-altered Gamma or the Tweedie model are better choice, but I would like to compare them to the lognormal model, i.e. glm (fml, family=gaussian(link="log"))).

The question is, can I add a constant to the DV and use this model? Also, do I just subtract the constant from the model estimates at the end? Is this valid at all?


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.