Centering of continuous variables before computing interaction terms in regression?

[Maria Sapouna]

Hi everyone,

Just another question that arised with regard to interaction terms in
regression. I would like to compute the interaction term of a continuous and
a categorical variable.
Some authors suggest that you just multiply the variables but in some others
texts I have read that you need to centre the continuous variable? If yes,
why?

Thank you in advance for all your help.

Maria