Regression ananlysis?

Question

In a regression analysis about the sales of a product corresponding to the week they were sold, why is it unwise to use the variables 'week' 'week squared' 'log week' and 'inverse week' in the same model?

Answer 1

It's not exactly unwise. In fact it could result in a truer interpretation of the data. It extends your matrix from a 2 x 2 to a 5 x 5 though, and it's a lot more tedious to extract coefficients. To use all these variables at the same time, you will have to use substitution variables.