Exploring asymmetric effects of attribute performance on customer satisfaction using association rule method. (May 2015)

• Record Type: Journal Article
• Authors: Chen, Li-Fei
• Is Part Of: International journal of hospitality management. Volume 47(2015)
• Page Start: 54
• Page End: 64
• Publication Date: 2015-05
• Subjects: Customer satisfaction -- Asymmetric effect -- Kano model -- Attribute performance -- Association rule; Hotel management -- Periodicals; Restaurant management -- Periodicals; Food service management -- Periodicals; Hôtels -- Gestion -- Périodiques; Restaurants -- Gestion -- Périodiques; Services alimentaires -- Gestion -- Périodiques; Food service management; Hotel management; Restaurant management; Periodicals; 647.94
• DOI: 10.1016/j.ijhm.2015.03.002
• Languages: English
• ISSNs: 0278-4319

Abstract

Highlights: The proposed framework successfully classifies Kano quality attributes. An AR method is used to evaluate the effect of AP on CS. The proposed method is more practical for classifying Kano quality attributes. The proposed method outperforms some regression-based approaches. Abstract: Identifying primary service attributes that generate customer satisfaction (CS) is critical to organizational success. The Kano model demonstrates an asymmetric relationship between attribute performance and CS. However, extant regression-based approaches for classifying Kano's quality attributes have theoretical limitations, such as multicollinearity problems, resulting in spurious inferences. The association rule (AR) method is widely used in data mining to explore the associations or correlations among variables because it does not require the typical assumptions associated with regression analyses. The framework developed in this study incorporates the AR method to classify Kano's quality attributes. The effectiveness of the proposed method was demonstrated using a case study of a restaurant chain. The proposed method is more practical for classifying Kano's quality attributes because it shortens the time required for data collection. Moreover, the proposed method reduces computational complexity. Validity test results indicate that the proposed method markedly outperforms some regression-based approaches.