Using every level of a categorical array in a regression

I am trying to run a regression, where instead of a constant term I would like to include every level of a categorical array. I know, that the results will be numerically the same either way, the main reason why I want to do this is convenience (it is part of a larger estimation procedure).

I would like to ask You if there is a way to do it when estimating a GLM model.

Thanks in advance,

Martin

Answer 1

[Tom Lane]

You could try using the DUMMYVAR function to generate indicator variables for each of your categories, then omit the constant term. If you supply the categorical variable directly as a predictor, MATLAB always omits one of the categories.