to anyone who can assist
consider the following lm model
dv ~ lm(a + a^2 + b + b^2 + a:b)
here i have fitted powers explicitly and fitted the interaction for the
*linear* components of a and b.
my question is how to i do the same thing using othogonal polynomials ??
i have tried
dv ~ lm(a(poly,1) + a(poly,2) + b(poly,1) + b(poly,2) +
a(poly,1):b(poly,1))
but i get an error message in splus of the sort:
Problem: Couldn't find a function definition for "a"
can anyone please tell me how to write the model in the orthogonal form
corresponding to the powers form. my aim is to be able to fit only the
linear components of a and b as the interaction term.
thanks
andrew
Dr AST Czarn
|