This cross sectional study tested the hypothesis that characteristics of work that contribute to job strain also increase anger in young service sector workers. A new measure of anger directed at co-workers, supervisors, and customers was regressed on job strain indices (job control, co-worker and supervisor support, dissatisfaction) in models that controlled for dispositional negative affect and work status. Results in a sample of 230 young Black and White men and women revealed that low levels of job control and social support, and high levels of job dissatisfaction, were independently associated with increased work-related anger. Moreover, social support moderated the impact of low job control on anger directed at co-workers. Findings indicate that anger experienced at work may be an early marker of job stress, which has been prospectively related to cardiovascular disease. We tested the hypothesis that youths who exhibit diminished social-emotional competence and frequent anger in adolescence experience increased occupational stress (with elevated coronary heart disease CHD risk) after they become employed. Measures of perceived job control and support from co-workers in 57 young Black and White men and women were regressed on measures of social problem-solving skill (SPS) and anger arousal that had been obtained 5 years earlier when participants were in high school. In models controlling for grade point average (GP A), SPS and GP A independently predicted co-worker support in young adulthood; anger in high school predicted diminished job control. These findings suggest that occupational stress, a risk factor for CHD, may have identifiable social-emotional antecedents early in life. A brief scale to measure emotional labor was adapted from the Emotions at Work Scale and evaluated in an ethnically diverse sample of working adolescents and young adults of both genders. As a result, a self-administered l3-item scale measuring emotional labor was refined to a 9-item scale. Five items loaded on the surface acting factor while 4 items loaded on the deep acting factor. The internal consistency of this scale demonstrated a mean inter-item correlation of 0.33 for each of the 2 factors. The overall Cronbach's alpha was 0.96. For the sub-scales, surface acting and deep acting, the Cronbach's alphas were 0.71 and 0.67, respectively. Test-retest reliability for the refined version was 0.64 for surface acting and 0.51 for deep acting indicating stability over a mean time interval of 3 months. In summary, this scale can be incorporated into other studies of the emotional labor experience among young workers.