| Emotion management, such as suppressing anger, is performed by individuals to cope in many social settings. Lab studies, however, have suggested that suppressing emotions or showing "fake" emotions requires physiological effort which may have stress outcomes over the long term. The implications of emotion management in customer service organizations, where employees are expected to display positive emotions to customers, are just beginning to be tested empirically. One difficulty in this area is that prior theorists examining "emotional labor," or emotional management for a wage, have not agreed upon either the definition of emotional labor or its measurement. The conceptualization of emotional labor proposed in this paper integrates previous theorists' ideas and applies lab-based emotion regulation theory to the field-based emotional labor topic. Emotional labor is conceptualized as emotion regulation to meet organizational demands, which can be achieved through surface acting (regulating expression) and deep acting (regulating feelings). Analyses of a measure of emotional labor also revealed a third scale which represented genuine expression of emotion, which was tested with post-hoc analyses. Hypotheses stating that organizational characteristics, such as frequency of interaction with customers and emotion display rules, would predict emotional labor were generally not supported. Hypotheses regarding the detrimental effect of emotional labor on burnout, job satisfaction, turnover intentions, emotional estrangement, and peer-rated customer service were generally supported. In particular, surface acting detrimentally predicted each of these outcomes beyond demographic variables, the organizational characteristics, deep acting, and genuine expression of emotions. Deep acting had a positive relationship with peer-rated service beyond the other variables, suggesting that some forms of emotion regulation are more beneficial than "faking" expressions. Implications of these relationships are discussed. Peer and supervisor support were tested as moderators of the detrimental relationships, providing limited evidence for a buffering hypothesis. Future research suggestions for this area of study are suggested. |
