TRAVEL BEHAVIOR: RESOURCE PAPER--PRODUCT ATTRIBUTES

A METHODOLOGY FOR ESTIMATING THE RELATIVE IMPORTANCE OF PRODUCT ATTRIBUTES IS OUTLINED AND ILLUSTRATED. PRODUCT-ATTRIBUTE DESCRIPTIONS ARE DEVELOPED IN TERMS OF ATTRIBUTE-SATISFACTION RATINGS OBTAINED BY A PARTICULAR TYPE OF SEMANTIC DIFFERENTIAL. THE USE OF SUCH RATINGS REQUIRED THE DEVELOPMENT OF A FAMILY OF DEMAND RELATIONS SPECIFIED IN TERMS OF ATTRIBUTE-SATISFACTION RATINGS. ATTRIBUTE IMPORTANCE WAS DETERMINED TO BE PROPORTIONAL TO THE MAXIMUM UTILITY OBTAINABLE FROM ANY LEVEL OF THE ATTRIBUTE. THE FAILURE TO SPECIFY EXISTENCE OF SUPPLY-SIDE RELATIONS AND FAILURE TO HANDLE SEMANTIC REDUNDANCY IN THE SET OF ATTRIBUTES, WILL CAUSE INCONSISTENT ESTIMATES OF IMPORTANCES AND VALUES WITH HIGH VARIANCE. IT IS SUGGESTED THAT DISCRIMANT ANALYSIS AND PRINCIPAL-COMPONENTS TYPE OF FACTOR ANALYSIS BE USED IN AN ITERATIVE FASHION TO DEVELOP A SET OF ATTRIBUTES THAT FULLY DESCRIBE THE PRODUCT FROM A CONSUMER POINT OF VIEW BUT ARE AS ORTHOGONAL AS POSSIBLE. THE DEMAND-SIDE RELATIONS WERE DEVELOPED ON THE BASIS OF 3 FUNDAMENTAL ASSUMPTIONS, NAMELY, ADDITIVE UTILITIES; EXPONENTIAL UTILITY FUNCTION SPECIFIED INDEPENDENT OF THE PRODUCT; AND LINEAR MAPPING UTILITIES TO ATTRIBUTE-SATISFACTION RATINGS. ESTIMATES OF RELATIVE IMPORTANCES VARY GREATLY DEPENDING ON MODEL SPECIFICATIONS. A METHODOLOGY IS SUGGESTED THAT CAN LEAD TO ESTIMATES THAT ARE BOTH EFFICIENT AND CONSISTENT. THE OBSERVATION IS MADE THAT THE IMPROVED UNDERSTANDING OF THE CONSUMER DECISION-MAKING PROCESS THAT WILL RESULT FROM THE IMPORTANCE-ESTIMATION MODEL DESCRIBED HERE, WILL OUTWEIGH THE TIME AND MONEY SPENT IN ITS DEVELOPMENT.