Full-profile conjoint analysis: some measuring, modeling and levels of aggregation

Abstract

In this work the various conjoint analysis (COA) models developed by the author are reviwed. Such models consider different levels of data response measurement scales; different levels of response aggregation, either individual or aggregated and different parameter estimation methods. We therefore report: i) some approaches to full-profile COA by multiple regression analysis: weighted least squares approach; the arcosin transformation approach; an additive binary coding of ordinal experimental factors; COA to estimate more than one response function; ii) some approaches to full-profile COA by multiple logistic regression analysis (ordinal logistic regression for the estimate of the response functions; multivariate logistic regression; multivariate logistic regression response with main and interaction effects).