Despite the study’s implications, several limitations must bead dressed. First, the sample consists of employees at family-style restaurants, part of the food service industry. Therefore, the gen-eralizability of the results may be limited to employees in those particular categories. In addition, the employees were selected based on their managers’ willingness to participate in this study.In the present study, the measurement items could have been influenced by the environment in which the employees answered the questionnaire; however, cross-sectional data could not be used due to temporal and spatial limitations. Therefore, the meaning of the results should be carefully determined. Analyses usedstructural equation models to examine the reliability and validityof the variables and verify hypotheses. The results indicated that the indexes theoretically presented by SEM did not reach the optimum goodness of fit. The goodness of fit of SEM cannot be judged based on only one index, and models are sometimes revised through revised indexes to enhance goodness of fit. However, the indexes were not revised in this study. Since all variables were measured through self-reporting questionnaire surveys, the results might have been distorted due to common method bias (CMB).Future research might also investigate personal variables, such as employees’ personal characteristics, job-related characteristics,and organizational culture, which could moderate the relationship samong the constructs. Also, this study was focused only on the negative effects of emotional labor as a prior factor of stress. Future research focused on the positive effects is therefore also necessary.