The perceived images of seven retail firms encompassing several lines of trade are presented in this paper. To determine these images, data on store attributes and the importance of these attributes were collected from a random sample within selected census tracts of a Southwestern city. The linear compensatory multi-attribute model approach of product attitudes was extended to that of store patronage. The ability of these image models to predict attitudes towards the retail outlets as well as patronage behavior is investigated. In addition, the power of importance weights in these models is statistically examined.