AN EXAMINATION OF THE PREDICTIVE ABILITY OF A MULTI-ATTRIBUTE ATTITUDE MODEL WITH USAGE SITUATIONS SPECIFIED

The purpose of this research is to examine the predictive ability of a multi-attribute attitude model under different usage situations. Model predictions of brand choice for three situations are compared to predictions by the model when usage situations are not specified. The product class employed for testing the model is facial tissue. Questionnaires were mailed in three waves to residents of the Greater Lafayette, Indiana area. A total of 565 usable responses was obtained. The data collected in this research quantify the respondent\u27s belief as to the extent which a particular brand holds a certain attribute. The data were examined by means of cross-tabulation and conditional probability tables. A proportionality test was applied to situation-specific data for tests of significant differences from nonsituation-specific data. The model correctly predicted first through third choices and eighth choice more often than fourth through seventh choices. Model predictions were consistent across situations. It was found that the multi-attribute model did not yield significantly better predictions of brand choice when usage situations were specified