regression

[poornima285]

could anybody tell me how to introduce interaction term in regression? say if the model is
y=a+bx1+cx2.
then how to introduce the term X1*X2? and then how to do the regression?

 

[royhaas]

It's just another variable: \(\displaystyle y = a +bx_1+cx_2+dx_1x_2\). Then use whatever program you normally use.

 

[poornima285]

thanks for your reply. suppose i hvae 13 predictors, with 12 of them being continous varibles and one being a dichotmous varibale, then on what basis should i introduce the interaction?