The aim of this paper was to examine whether characteristics of the assessed actor and empathy-inducing instruction presented to the participants influence self-reported empathy as well as self-reported behaviour (willingness to help and assigned length of imprisonment) and moral reasoning. The research was conducted on the sample of N = 206 participants, college students from the University of Osijek. However, data from 11 participants was not processed, leaving a sample of N = 195, from which 129 were female (66.2%). Participants were randomly divided into 4 experimental situations in which they read different versions of a scenario in which the actor committed a theft. The scenario was created for the purpose of this research. Ethnicity of the actor (Croat or a Romani) and presented instruction (with or without empathy induction) were manipulated in a scenario. Afterwards, participants were supposed to answer questions concerning the length of imprisonment they think the actor deserved and to which extent they would be willing to help out his family. Empathy was measured by the adapted State Empathy Scale (Shen, 2010), and moral reasoning was measured by the adapted Test of Moral Reasoning (Proroković, 2016). The results indicate no effect of the actor’s ethnicity in a scenario, empathy-inducing instruction nor their interaction on the self-reported empathy, willingness to help, assigned length of imprisonment and moral reasoning. Additionally, the results suggest that there is a significant difference in self-reported empathy with respect to gender and place in which participants grew up. Regarding participants' field of study, there is a significant difference in both empathy and willingness to provide help. Furthermore, results show that participants who were presented with the scenario in which the actor was Romani evaluated Roma people significantly more positive than those who were presented with the scenario in which the actor was Croatian.