Date: Nov 21, 2012 11:13 PM
Author: Halitsky
Subject: Interpretation of coefficients in multiple regressions which model<br> linear dependence on an IV

In a different thread, Ray Koopman explained that if one suspects
these regressions to be dependent on the IV ?u?:

c on u
c on e
c on (e,u)

then under the usual initial assumption that the dependence is linear,
these three regressions should be modified to:

c on (u,u^2) instead of c on u
c on (e, u, u*e) instead of c on e
c on (e, u, u*e, u^2) instead of c on (e,u)

Ray also indicated that proper intepretation of coefficients in such
regressions is a topic deserving of its own thread.

And therefore, I have made this post to start a thread that Ray can
use whenever he has time and inclination to begin his exegesis.

But I do have one ?dumb? question to ask Ray before this exegesis
begins.

For multiple regressions, I have been using Ivo Welch?s PERL
Statistics::Regression package and it?s been working fine.

But when I give his package the regression

c on (u,u^2)

it complains that it can?t compute a covariance matrix for the
coefficients.

Is this to be expected because the one IV is the square of the other ?

Or am I doing something else wrong here?