User:Georges/Standardize or Not
John Fox's good example of the problem with standardization:
Consider a regression of income on years of education and on years on experience. Suppose the coefficient for years of education is $2,000 and the coefficient for years of Experience is $1,000. Then we can compare their relative value and consider the income predicted with an additional year of education and, consequently, a year less of experience at a given age. However, if we first standardized with, using a typical population, a much larger sd for experience than for education, the comparison of the coefficients would be meaningless.