I am regressing meat consumption (a continuous variable) on several socio-demographic characteristics (all categorical variables) converted into dummies : sex (1/2), age category (1/2/3), level of education (1/2/3/4), size of city (1/2/3/4/5), revenue category (1/2/3/4) and number of children (0/1/2/3). For the regression, I remove one dummy of each socio-demographic variable, which is my reference individual.

My problem is that the coefficient of each dummy (female for example) depends a lot on the reference individual chosen. Is there a specific method to compute the absolute effects ? For example, the absolute effect of "being a woman", controlling for all other socio-demographic variables.

I am using STATA, is there a specific command ?

Thank you very much, and sorry if the question have already been answered but I searched for quite some time !



Your Answer

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