Mean centering in regression in SPSS

Mean centering predictors in regression Recognised reasons for mean centering continuous IVs on higher order terms such as interactions, polynomial terms (X to the power of 2 for example): aids interpretation of the intercept term - it's the mean of DV when all Xs are set to their mean values numerical stability in computing estimates (but not such an issue) there is debate whether one mean centers to reduce multicollinearity. Firstly, if you don't have interaction terms, polynomial terms, then mean centering does not deal with multicollinearity. Secondly, supposing you had interaction or polynomial terms, then so long as you include all the lower order terms, you get the same estimate, se on the coefficient on the interaction/ polynomial term. The coefficients on the smaller terms will change as the coefficients on centered, and non centered Xs carry different meanings. So here mean centering is not a way to reduce multicollinearity.