The on-going (self-)monitoring of our behaviour is inextricably intertwined with the surrounding social context. In this study, we investigated whether a minimal group paradigm assigning individuals to arbitrary group categories is powerful enough to induce changes in behavioural, psychophysiological and event-related potential correlates of performance monitoring. Following arbitrary group assignment based on ostensible task performance and a group identification task, 22 volunteers performed a flanker-task during both in-group and out-group contexts, while electroencephalography was performed. More errors were committed in the out-group compared with the in-group context. Error-related negativity amplitudes were larger for in-group compared with out-group errors. However, subsequent processing reflected in late Pe amplitudes and stimulus-driven conflict reflected in N2 amplitudes were not affected by the group context. Heart rate deceleration (during both correct and incorrect trials) tended to be more pronounced during the out-group compared with the in-group context. This surprising observation was corroborated by subjective ratings of performance satisfaction, in which participants reported higher satisfaction with their out-group performance. This study identified specific stimulus evaluation processes to be affected by a minimal group manipulation and demonstrated thereby transient top-down effects of a social context manipulation on performance monitoring.