Thread: Help with multivariate regression model

Hey everyone,

I am currently working on a creating a model that is a function of three variables:

Performance = 0.xx (Net Cost) + (0.xx) Best Practice Score - (0.xx) Net Tonnes marketed per household.

Seems simply enough right?

The problem is, is that before I even assign weights to each respective covariate, the model is heavily scewed in favor of net cost as the actual number itsef is significantly larger than both Best practice score and net tonnes per household.

What could I do to make sure that all the variables are of equal importance (with respect to its effect on performance), before I assign the probability weights for 0.xx?

Any help would be greatly appreciated!

If you are just fitting a model with least squares, how you scale the predictors is completely irrelevant. If you scale a predictor up by a factor of 10, you are just going to see the associated slope decrease by a factor of 10. You can't force the predictors to be of "equal importance."

If I've completely missed the point, it would help if you were a little more explicit with your notation and with what exactly your problem is.