Multiple Regression

[rodcamb]

hey,

i want to estimate a model for my dissertation in the form of

y= a+bx+cz+e

where a,b,c are coefficients and x and z are independent variables.

I know I could use OLS, but would like something more advanced than that, what sort of options do I have? or should more advanced than ols be down to any errors etc that i encounter?

[Carlo Lazzaro]

Rodcamb:
assuming you have one wave of data only and the relationship between depvar and indepvars is linear, OLS is the way to go.
Obviously you can add square-terms (or higher) if they fit your data better, as in OLS the assumption of linearity concerns coefficients, not variables.