Improving the Predictive Validity of Quality Function Deployment by Conjoint Analysis: A Monte Carlo Comparison

Abstract

Recently, for quality function deployment (QFD), new approaches have been developed to improve the estimation of the influence of technical features and engineering characteristics on relevant product attributes and of the importance of these attributes for the customer. Instead of group discussions with qualitative consensus outcomes these new approaches make use of regression analysis or of conjoint analysis. Applications indicate superior predictive validity of these new approaches. In this paper, this superior validity is analyzed in more detail. A factorial design with synthetically generated data is used for comparisons.