Introduction: Previous studies on readiness to forgive have rarely linked punishment and forgiveness or have observed them separately. Although these two constructs and processes may seem mutually exclusive, both play a role in establishing social balance (Strelan et al., 2013). Readiness to forgive is an important component in resolving many interpersonal conflicts, and punishing the offender as a possible response to an offense with the aim of restoring social balance can have a positive effect on the victim's recovery. Therefore, the aim of this study was to examine the relationship between certain circumstances of the offense (the level of personal responsibility of the offender, the expression of regret by the offender for the committed offense) and the anger of the evaluator with the assessment of the punishment for the offender and the degree of readiness to forgive towards the offender. In the forgiveness process, responsibility has been identified as an important factor in some studies, where lower offender responsibility is associated with a higher likelihood of forgiveness (Eaton and Struthers, 2005; McCullough et al., 2003). Furthermore, according to previous research, the regret of the offender also increases the likelihood of forgiveness, as the offender expresses regret for the inflicted harm, eliciting empathy from the victim and observer of the offense (Scher and Darley, 1997; Struthers et al., 2010). The inflicted harm triggers various emotions in the victim and the observer, with anger being one of the most common. In this study, the commission of the offense is viewed as a stressful event that elicits emotional reactions from the observer. According to emotion research, anger can enhance the desire to punish the offender (McCullough, 2000) and, on the other hand, can decrease the possibility of forgiveness (Younger et al., 2004, as cited in Merolla, 2008). To our knowledge, this is the first study in which the responsibility and regret of the offender, as well as the anger of the evaluator in assessing the punishment and readiness to forgive, were simultaneously examined. Methodology: The participants in the study were undergraduate and graduate students from various faculties of three Croatian universities. A total of 587 participants took part in the study, including 188 males (32%) and 399 females (68%). The research was conducted in groups during lectures. Participants were randomly assigned stimulus materials related to different experimental situations. The study used scenarios describing the same criminal act (a traffic accident resulting in a fatality) with variations in the circumstances of its commission: the offender's level of responsibility (two levels: high and low) and the offender's expression of regret (two levels: regretful, not regretful). The research design included three different contexts, and within each context, four described stimulus scenarios were used. In the first context, the evaluator had no information about legal sanctions against the offender. In the second context, the evaluator had information about the sentenced punishment for the offender, while in the third context, the evaluator had information that the offender was not legally sanctioned for the committed offense. Regardless of the group they belonged to, all participants read a description of the criminal act and then provided two types of assessments. First, they assessed how much the described act evoked anger in them, using a scale from 1 to 10. After the anger assessments, participants evaluated the appropriate punishment for the committed act and their readiness to forgive the offender. Before providing these assessments, participants were exposed to one of three different pieces of information regarding the legally determined punishment for the offender. Participants expressed their punishment assessments as the number of years of imprisonment they considered appropriate for the committed criminal act, regardless of existing legal regulations, and readiness to forgive was measured using a 16-item Forgiveness Scale, with the average result on the questionnaire used as a measure of readiness to forgive. Results and Discussion: Given the factorial research design of 2x2x2, three way ANOVA were conducted, two for each context, with independent variables being the degree of offender responsibility, the existence of offender regret, and the degree of evaluator anger, while dependent variables were the assessment of punishment for the offender or the degree of readiness to forgive. Although it was expected that participants would give a harsher punishment and be less ready to forgive at a higher level of offender responsibility and give a milder punishment and be more ready to forgive at a lower level of responsibility in all three contexts, the main effect of offender responsibility was statistically significant only in the context where the offender was not punished. In this context, participants, knowing that the offender was not legally punished, gave harsher punishments to the more responsible offender than to the less responsible offender. It is possible that in other contexts, the main effect of offender responsibility was not observed because the responsibility variable was not manipulated convincingly enough. Since the outcome of the event at both low and high levels of responsibility was associated with a fatal outcome, it is possible that the death of the victim determined responsibility, and participants, in contexts where they either knew that the offender was punished or had no information about the punishment, assigned punishments regardless of whether they were exposed to a scenario of low or high responsibility. However, the question of manipulating the responsibility variable still does not explain why a significant main effect of responsibility was obtained in the context where participants knew that the offender was not punished. This could be explained by the fact that in the context where the offender was punished, participants did not take responsibility into account because justice was already satisfied by the fact that the offender was punished. Regarding responsibility and readiness to forgive, the main effect of offender responsibility was not significant in any context. Besides the explanation that the responsibility variable was probably not manipulated clearly enough, the obtained results, indicating that offender responsibility was not relevant to readiness to forgive, can be explained by an additional methodological aspect of our study. Namely, participants assessed their readiness to forgive the offender for an offense committed against another person. Comparing our study with other studies, such as the one conducted by Struthers and colleagues (2005), where participants imagined that the event happened to them personally, it can be noticed that personal harm was important for readiness to forgive, which was not the case in our study with this scenario nor was it possible to achieve it with this scenario. In addition, participants did not assess the level of offender responsibility themselves; rather, responsibility was a priori defined by the content of the vignette, questioning how successfully it was manipulated. The factor of offender regret is relevant in all three contexts for both forgiveness and punishment, with regret leading to lower punishments and greater readiness to forgive. Regret as the main effect in the assessment of punishment and readiness to forgive in all three contexts always operates in the same direction, with participants giving lower punishments when the offender expressed regret compared to when the offender did not, and they were more ready to forgive when the offender expressed regret, while they were less ready to forgive when the offender did not. The obtained findings regarding the punishment assessment align with previous knowledge, suggesting that if the offender suffers, the readiness to punish the offender is likely lower. According to previous studies, regret is important after a harm because it reduces the possibility of revenge or punishment of the offender, which are often responses to the inflicted harm (McCullough, 2008). Regarding readiness to forgive, the results are in line with previous research (Scher and Darley, 1997; Watanabe et al., 2020), indicating that participants were more ready to forgive the offender when the offender expressed regret compared to situations when the offender did not express regret. Also, anger is an important factor in punishment and forgiveness in all contexts, and it is shown that angrier participants generally administer more punishment and are less ready to forgive, except in the context where participants knew that the offender had been punished. In that context, anger had no effect on the assessment of the magnitude of the punishment. In the assessment of punishment in the remaining two contexts, participants who were angrier at the offender also gave a higher punishment compared to those who were less angry, as expected. This finding is in line with the theoretical assumption that emotional reactions such as anger and resentment, which occur after a harm, motivate a person to seek revenge and punish the offender (McCullough, 2000). Although the harm occurred, the victim's reactions will depend on the offender's subsequent behavior (e.g., whether the offender regrets their actions). Regret after a harm has been shown to be a factor that diminishes the victim's anger towards the offender, increases the victim's perception of the value of the relationship, and contributes to forgiveness after a harm (McCullough et al., 2014). Research clearly shows that the emotion of anger reduces the tendency to forgive. While the independent operation of the examined variables in selected contexts is more or less consistent, it is important to consider their interactions in interpreting their effects. Thus, for example, there is an interaction between responsibility and anger in the assessment of punishment in the context where participants had information that the offender had been punished, and in the same context, there is an interaction between the degree of offender's responsibility, the offender's regret, and the assessor's anger in the assessment of readiness to forgive. Furthermore, there is an interaction between regret and anger in the assessment of punishment in the context where participants had information that the offender had been punished and in the context of readiness to forgive when participants had no information about the punishment of the offender. In the end, it was found that participants were least ready to forgive in the context where the offender was not punished for the offense, and forgiveness assessments in this context were significantly lower compared to contexts where participants had no information about whether the offender was punished or not, as well as the context when they had information that the offender was punished. Our results are in line with previous studies (Strelan et al., 2013, 2017) indicating that participants were more ready to forgive offenders if they were punished. Conclusion: In this study it was shown that the main effect of responsibility was statistically significant only in the context when the offender was not punished. The regret of the offender is relevant in all contexts, leading to lesser penalties and greater readiness to forgive. Anger played a role in both punishment and forgiveness, except when participants knew that the offender had already been punished. Overall, responsibility, regret, and anger play crucial roles in the assessment of punishment and readiness to forgive after an offense. Although the independent effects of the examined variables in selected contexts are more or less consistent when interpreting their actions, it is important to consider their interactions, which are context-dependent, and their effects are interdependent for forgiveness, only in the context where participants knew that the offender had already been legally punished. In that context, participants who were differently angry at the offender were variably ready to forgive depending on whether the offender had expressed regret and whether he was more or less responsible for the accident: when the offender was more responsible, angry participants were equally ready to forgive regardless whether he expressed regret or not. However, in the case of less angry participants, regret was crucial, and those who received information that the offender regretted were more ready to forgive. On the other hand, in situations of low offender responsibility, the reactions of differently angry participants were precisely the opposite: now angrier participants were more ready to forgive with the regret of the offender than when the offender did not express regret, while less angry participants were ready to forgive equally, regardless of whether the offender expressed regret or not. Within the framework of the second research problem, it was found that participants were least ready to forgive when the offender was not punished, with forgiveness ratings significantly lower than in situations where participants did not have information about punishment or knew that the offender had been punished, which is consistent with previous studies.