centered polynomials

in a model involving Y= B0 + B1X1 + B2X2^2,

how does centering the model affect the parameters.

why does the hypothesis test associated with B2 stay the same, regardless of centering?


why does the hypothesis test involving both B0 and B1 change?



Ambassador to the humans
Centering helps with an issue called multicollinearity.

If you want to see why it doesn't have an influence on B2 but it does on B1 and B0 try writing out the model and expanding the centered terms to see what happens.