A service recovery performance model is proposed and tested with data from frontline bank employees in Turkey. The model is derived from Bagozzi's (1992) reformulation of attitude theory. The empirical results suggest that top management commitment to service quality, as manifested by frontline employees' appraisal of training, empowerment, and rewards, has a significant effect on their perceptions of service recovery performance. The influence of management commitment to service quality on service recovery performance is mediated by frontline employees' affective commitment to their organization and job satisfaction. Implications of the results and further research avenues are discussed.

According to prior researches about service failure and service recovery, much attention was given to types of service recovery. However, there was still lack of discussion on the effectiveness of service recovery performance on customer satisfaction, WOM and re-patronage. Recently, there were some researches investigated on how customer relationship with a service organization affects their reactions to service failure and recovery. As a result, this research integrated the five determinants of customer satisfaction (stability attribution, controllability attribution, perceived equity, positive and negative affect) to further investigate the moderate effect of customer relationship on service recovery performance, WOM and re-patronage. This research reports an experimental design in the context of restaurants is separated into (2*4) eight scenarios. Four hundred and twenty-two questionnaires were gathered via internet. According to the gathered data, the conclusions of this research are summarized as following: 1.As the service recovery performance increases, customers are unlikely to attribute the service failure as a stable factor. At the same time, the degree of customers¡¦ perceived equity would be higher. 2.Customer¡¦s relationship is significant to the stable attribution, but is not significant to the controllable attribution and perceived equity. 3.Negative relationships exist between stable attribution, controllable attribution and positive affect. Positive relationships exist between perceived equity and positive affect. 4.Customer¡¦s relationship is not significant to moderate the re-patronage and word-of mouth. 5.The integrated analysis on cognitive and affective viewpoint can let post-recovery satisfaction clarify 66.2% of variation explanation. 6.The LISREL analysis result appears that the goodness of fit of the structure in this study would be acceptable.