Under what circumstances would you expect income to be a better predictor of consumer behavior than a

Question:

Under what circumstances would you expect income to be a better predictor of consumer behavior than a composite measure of social class (e.g., based on income, education, and occupation)? When would you expect the composite social class measure to be superior?