Research suggests that music can affect mood and physiological states (Elliott, Polman, & McGregor, 2011;Thoma et al., 2013). The extant literature has mainly focused on long-term changes in self-report measures resulting from several sessions of music intervention (but see Jiang, Zhou, Rickson, &Jiang, 2013 and Pelletier, 2004 for review). The purpose of this study was to expand the previous work in three major ways. First, the stress-reducing effects of music listening versus musical improvisation on a xylophone were compared. Second, psychophysiological measures of stress were utilized in addition to self-report measures. Third, stress reduction effects were measured in one session. A total of 105 participants completed a stressor task and were then assigned to one of three recovery conditions: control (CG), music listening (ML), or music improvisation (MI). The ML and MI groups were expected to exhibit greater stress reduction (evidenced by behavioral and physiological measures) than the CG. Self-report and skin conductance (electrodermal activity; EDA) measures were used to assess response during the stressor task and recovery session. Results indicated that participants exhibited a significant increase in stress response as a result of the stressor task. Although self-report measures indicated that mood improved during the recovery session, the MI group exhibited only differential effects for levels of satisfaction. The physiological data showed greater stress reduction for the ML condition as compared with the CG and MI group.