Working towards social justice necessitates dominant group members’ willingness to share the feelings of oppressed outgroup members (outgroup empathy) and to engage in difficult intergroup dialogue. Nevertheless, empathic and behavioral engagement is especially avoided in social justice contexts due to the high demands/costs of such engagement. It is, however, possible to offset high demands and motivate an approach-oriented response by increasing resources, namely social resources. Compassion trainings have been shown to expand social resources by increasing perceived social connection. This study investigates (1) the effect of a two-week compassion training (vs. active control) on White participants’ prosocial engagement (i.e., outgroup empathy, approach/avoidance physiology, and willingness to engage in social justice dialogue) in a situation where race-related social injustice is highlighted; and (2) whether such effects occur through an expansion of perceived social resources. Participants’ prosocial engagement and perceived social resources will be measured before and after training in a novel laboratory paradigm involving behavioral, physiological, and self-report assessments of prosocial engagement. This work will have important practical and theoretical implications by advancing our understanding of the processes through which compassion training shapes prosocial outcomes in a social justice context that is highly relevant to today’s society.