Componential Segmentation Based Conjoint Analysis: the Dummy-Coded Parametrization of the Model

Amedeo Pasquale De Luca

Risultato della ricerca: Contributo in rivistaArticolo in rivistapeer review


In Componential Segmentation (CS) each component of each consumer profile, considered jointly with each component of the product profile, is supposed to contribute to the overall evaluation of the product. The CS focuses on the effect of interactions between the product profile x (a vector of dummy variables that describes the product) and the person profile z (a vector of dummy variables that describes the person in terms of a certain set of background characteristics) on preference for the product. A consumer’s reaction to a product is broken into the sum of two components: 1) the average part -worth utilities due to the attribute levels of the product (pooled across all respondents in a market survey) and 2) the interactions between the consumer’s background variables and the attribute levels. In this paper we propose the dummy-coded parametrization of the model, easier to apply than other types of estimation suggested in the literature, and which provides two baselines (a new circumstance in the literature of the field): a first one related to the attributes of the product, a second one relating to the background characteristics. We provide an application of the model and an original interpretation of interaction effects. © 2014 The Authors. Published by Elsevier B.V. Selection and peer-review under responsibility of the Organizing Committee of IES 2013. Keywords: Baselines; Componential Segmentation; Conjoint Analysis; Decompositive Model; Interaction effects; Dummy variables.
Lingua originaleEnglish
pagine (da-a)175-182
Numero di pagine8
Stato di pubblicazionePubblicato - 2014
Pubblicato esternamente


  • Baselines
  • Componential Segmentation
  • Conjoint Analysis
  • Decompositive Model
  • Dummy variables.
  • Interaction effects


Entra nei temi di ricerca di 'Componential Segmentation Based Conjoint Analysis: the Dummy-Coded Parametrization of the Model'. Insieme formano una fingerprint unica.

Cita questo