A Statistical Thinking Approach to Kansei Engineering for Product Innovation
Tarantino, Pietro (2008) A Statistical Thinking Approach to Kansei Engineering for Product Innovation. [Tesi di dottorato] (Inedito)
Full text disponibile come:
With a strong competition and a strong consumer awareness of quality, manufacturers have to look hard at how to satisfy needs and expectations of potential new consumers. The only acceptable level of quality is total. In addition to the functional needs, affective and emotional needs have been recognized as having primary importance for consumer satisfaction and for creating innovative products. Kansei Engineering is a newly emerged product development technique to deal with consumers’ feelings and emotions and to incorporate these emotions into design elements during the product concept design phase.Kansei Engineering has enormous potentiality, nevertheless to be successful and really innovative, it needs to be integrated with the traditional methodologies for product design and to be supported by quantitative methods. The underlying aim of this research work is to minimize intuition in design decisions and to maximize the systematic use of statistical methods in product concept design phase. These methods can provide design team with the analytical tools for correctly plan experimental phases in Kansei Engineering and for analyzing the results in a reliable and efficient way. In particular, the advancements in Kansei Engineering and product concept design methods that this research has attempted to bring about were developed through five research mainstreams. The first research mainstream aimed at formalizing an integrated approach for incorporating both functional and emotional quality elements in product concept design. The proposed approach makes use of statistical methods, such as supersaturated design and ordinal logistic regression, for product concept arrangement and consumer data evaluation, while contemporarily emphasizing the use of virtual reality technology for consumer-designer interaction. Secondly, despite the large literature on the use of design of experiments for a statistically valid formulation of product concept, few works in the Kansei Engineering area make use of such tools. Therefore, the second research mainstream aimed at suggesting the most efficient design for Kansei Engineering experimentation. In particular the properties of saturated and supersaturated design are explored. The third research mainstream aimed at introducing a general methodology for filtering the biasing effect of global noise factors on consumers’ evaluation. These noise factors arise when real products -taken from market- are presented to consumers in place of physical or virtual prototypes. The proposed methodology is tested for applications in Kansei Engineering, as well as for marketing and medical research areas. The fourth research mainstream aimed at providing statistical evidence of the goodness of non-linear models such as Ordinal Logistic Regression and Categorical Regression for data coming from a Kansei Engineering experimentation. Moreover, differences between rating and ranking procedure are analytically explored. Lastly, since the predominant research paradigm in product concept design is to consider a product as a bundle of well-defined attributes, an innovative methods for estimating consumers’ attribute importance is discussed. It allows to overcome most of the problems with context, survey and cognitive variables, since it uses an indirect procedure hiding the true task to the respondent of a controlled interview. A multidisciplinary approach, with knowledge from cognitive psychology, behavioural science, psychometrics, consumer research and marketing science is throughout used. Moreover, this research was stimulated by practical needs and always considering statistics as the fuel for the engine of innovation. Most of the contributions in this thesis are, in fact, validated through case studies carried out in a strong-collaboration with industrial designers and final consumers.
Solo per gli Amministratori dell'archivio: edita il record