Offense Type as Determinant of Revenge and Forgiveness After Victimization: Adolescents’ Responses to Injustice and Aggression

ABSTRACT

Victims of injustice and aggression may have strong feelings about the perpetrator(s) that may impede their efforts to cope with the victimizing experience. We examined to what extent adolescents’ interpersonal responses to victimization in terms of revenge and forgiveness depend on offense type. Of 455 Dutch students from various educational levels, 379 participants reported being victimized by incidents of injustice, aggression, or violence. These incidents were categorized according to type and related to respondents’ self-reported revenge, avoidance, and benevolence toward the perpetrator. Victims of criminal offenses (physical and sexual violence, theft, and threat) reported less forgiving motivations than victims of noncriminal transgressions (bullying, ostracism, and other forms of indirect aggression). Sexual violence primarily elicited avoidance, rather than revenge. Gender differences in responses to victimization depended on offense type, too. Hence, to enhance our knowledge about revenge and forgiveness after victimization, future studies may need to take offense characteristics into account.

KEYWORDS:

• Aggression
• avoidance
• benevolence
• forgiveness
• revenge
• transgression related interpersonal motivation (TRIM)
• victimization

Introduction

Victimization is a prevalent experience in many adolescents’ everyday life (e.g., Finkelhor, Turner, Shattock, & Hamby, 2013). The risk of becoming the victim (or perpetrator) of an act of injustice or aggression peaks in adolescence (e.g., Cops & Pleysier, 2014), and correlates with impaired mental and psychosocial development (e.g., Boney-McCoy & Finkelhor, 1996). As those authors have cogently argued, violence toward youths is a prevalent problem that can take many forms, all of which have the potential to disrupt the developmental process. One such disruption is in school performance such as academic and intellectual underachievement, commitment to learning (Hoglund, 2007; Shonk & Cicchetti, 2001), and school drop-out (Gary & Campbell, 1998; Harris, 1983). Research findings show that victims also have a high risk of multiple victimization (e.g., Turner, Shattuck, Finkelhor, & Hamby, 2015), ameliorating the developmental risk (e.g., Turner, Shattuck, Hamby, & Finkelhor, 2013). Moreover, victims may become offenders in turn, as studies on cycles of violence have shown for aggression in general (e.g., Dodge, Bates, & Pettit, 1990; Stillwell, Baumeister, & Del Priore, 2008; Widom, 1989), and for school violence in particular (e.g., Leary, Kowalski, Smith, & Phillips, 2003).

Victimization concerns violations of legal or moral rules that can be conceptualized as interpersonal transgressions (i.e., acts that are either morally wrong or offensive to one of the relationships partners, or inflict that partner psychological or physical pain; Worthington & Wade, 1999). This conceptualization refers to a wide variety of experiences, including relatively common offenses like bullying, lying, cheating, and gossiping, as well as less common legal offenses like robbery and physical and sexual violence. A common denominator is that victims have to cope with an adverse experience, and that part of their coping effort concerns their thoughts, feelings and actions towards the perpetrator(s): While some victims retaliate, others forgive or forget. The dynamic processes underlying the question why some victims take revenge while others forgive or forget constitute a complex puzzle that involves the interplay between characteristics of the victim, the perpetrator, the offense, and their context. This study addresses a small piece of that puzzle by investigating to what extent the motivation to forgive or pay back after being harmed by another person depends on the type of the offense.

Coping with victimization involves a variety of challenges, and one’s thoughts, feelings, and actions toward the perpetrator are but one dimension in the process. Within this dimension, the urge to retaliate and take revenge (i.e., an aggressive response to a person or to persons who inflicted oneself or one’s family intentional harm; Stuckless, 1996), is a common but not inevitable response. In the literature, a variety of alternative responses are mentioned; for instance, tolerate the harm and refrain from responding to the perpetrator, identify with or submitting to the perpetrator, bond together with other victims, express disapproval in an assertive way, follow official channels to attain justice, ignore or trivialize the transgression, avoid the perpetrator and all thoughts of the transgression, find a meaning that transcends the apparent pointlessness of the transgression, try to understand and sympathize with the perpetrator, and put the pain behind them and forgive the perpetrator (Frijda, 1994; McCullough, 2008; Schumann & Ross, 2010). Indeed, Strelan and Wojtysiak (2009) empirically demonstrated how coping strategies were related to the forgiveness process, bolstering the conceptualization of forgiveness in terms of coping.

McCullough and colleagues (McCullough et al., 1998) proposed a three-dimensional model to represent the variety in interpersonal responses to transgressions. According to this model, victims’ responses towards the perpetrator can be described in terms of their avoidance, and their vengeful or benevolent approach of the perpetrator. These three transgression-related interpersonal motivations (TRIMs) together provide an index of the extent to which one is inclined towards forgiveness. Forgiveness, conceived as a change process towards more positive and less negative thoughts and feelings about an individual who inflicted one intentional harm (McCullough, Root, & Cohen, 2006), is inferred from low revenge and avoidance motivation combined with high benevolence (McCullough et al., 1998).

As was delineated in Worthington and Wade’s (1999) model of unforgiveness and forgiveness, the question whether one responds to a transgression with revenge, avoidance, or forgiveness presumably depends on an interplay of intrapersonal (e.g., gender (Fehr, Gelfand, & Nag, 2010; Miller, Worthington, & McDaniel, 2008), narcissism (Exline, Baumeister, Bushman, Campbell, & Finkel, 2004), interpersonal (e.g., quality of the victim–offender relationships (McCullough, 2008)), as well as contextual factors (e.g., whether one lives in a community where justice can and will be restored (Schumann & Ross, 2010)). While these factors may all explain part of the variance in TRIMs, some offenses may simply be harder to forgive or forget than others (McCullough, 2008; Rapske, Boon, Alibhai, & Kheong, 2010; Worthington & Wade, 1999). To illustrate, a survey by Van Biema, Cole, Mitchell, Monroe, and Laughlin (1999) showed that 67% of the participants said they were able to forgive the thief of their money, while 15% thought they could forgive the murderer of their child.

Despite its intuitive appeal, the hypothesis that offense type affects feelings of revenge and forgiveness has as yet scarcely been tested (Carmody & Gordon, 2011). Studies in this realm have primarily focused on perceived severity of the offense: Offenses that are perceived as more severe are more likely to be avenged and less likely to be forgiven (e.g., Bradfield & Aquino, 1999; Fincham, Jackson, & Beach, 2005; Rapske et al., 2010). However, perceptions of offense severity are essentially subjective. They constitute a “fuzzy reality” (Konrath & Cheung, 2013) that blends characteristics of the offense with characteristics of the victim:

If given the choice between a broken bone and being dumped by a romantic partner, or between a black eye and being slandered by a close friend, many would seriously consider enduring the physically painful options over those that are psychologically and relationally so. (Barnes, Brown, & Osterman, 2009, p. 400)

Others would not, and adding a financially painful option would, for some people, complicate the equation even further. These dilemmas illustrate that perceived severity is a unidimensional index of a qualitatively varied experience: It collapses qualitative differences between offense types into a single score. On the other hand, offense characteristics are undoubtedly related to perceived severity in the sense that some offenses are perceived by most people as unforgivable such as, for example, sexual violence and murder (Rapske et al., 2010). These considerations point to the possibility that qualitatively different offenses elicit different TRIMs towards the perpetrator. Emotions that are elicited by the offense and the victims’ subsequent emotion contingent cognitive processing may play a mediating role here (Wenzel, Okimoto, Feather, & Platow, 2008). For instance, being robbed of a valued possession or being insulted or hit may predominantly evoke anger, and increase the risk of aggressive retaliation (Denson, 2013), whereas being sexually assaulted is more likely to elicit posttraumatic stress reactions like shock, fear, and confusion (Herman, 1992), with concurrent internalizing thoughts (e.g., self-blame, lowered self-esteem; Foa & Riggs, 1994) and feelings (e.g., depression, suicidal ideation; Regehr, Alaggia, Dennis, Pitts, & Saini, 2013). Typically, rape survivors are inclined to avoid all memories of the traumatic event (Jaycox, Zoellner, & Foa, 2002). Ostracism may initially activate the victim’s need to belong, driving them to ruminate about their own wrongdoings in order to discover ways to make amends (Williams, 2007). In sum, on the lead of emotions, victims may gravitate towards avoidance (e.g., in fear, disgust, disdain, and self-conscious emotions like shame, guilt, and embarrassment), benevolence (e.g., in need to belong), or revenge (e.g., in anger; Worthington & Wade, 1999).

Empirical research on qualitative offense characteristics as determinants of revenge and forgiveness appears sparse. Rapske et al. (2010) categorized undergraduates’ recollections of interpersonal transgressions that they were unable or unwilling to forgive, and concluded that unforgiven transgressions varied considerably in type and severity. Crombag, Rassin, and Horselenberg (2003) asked college students what had prompted them to take revenge, and found no differences in revenge motivation for different types of offenses. However, their participants reported transgressions that involved indirect aggression (i.e., false accusations, 21%; violation of trust, 21%; deserting one when help is needed, 13%; making one look foolish, 11%; or lying about something important, 9%). None of these categories involved direct aggression and violence, sexual violence, or some other kind of legal offense. Hence, the conclusion that type of offense was unrelated to revenge motivation may have been due to restriction of range.

Study rationale and objectives

The extent to which adolescent victims tend to avenge, avoid, or forgive a transgressor may affect their social, emotional, and academic development. In the worst case, victimization and ensuing vengeful rumination may initiate a cycle of violence that transforms the original victim into a perpetrator. For instance, McCullough, Kurzban, and Tabak (2013) reviewed that revenge was identified as a causal factor in 10%–20% of the homicides worldwide, 61% of school shootings, 27% of bombings, and an inspiring factor for recruitment in terrorist organizations. Cognitive models of aggressive offending also identified revenge as a primary motive for violent offending in general (Denson, 2013), for fire-setting (Barnoux & Gannon, 2013), sexual violence (Barnett, 2011), mass murder (Scheff, 2011), and (suicide) terrorism (Lankford & Hakim, 2011). According to these formulations, victims may, on the basis of their adverse and sometimes traumatic experiences, develop schema’s or implicit theories that define the world as a place of danger and injustice, which affect the perception and processing of new social information and justify the use of retaliatory action. Moreover, harboring grudges impedes healthy coping with trauma (e.g., Kunst, 2011; Orth, Montada & Maercker, 2006). In sum, revenge motivation might set the stage for an adverse developmental trajectory. Avoidance of the perpetrator and avoidant coping with victimization may at first sight seem less (violently) harmful, but appears adversely related to positive and healthy outcomes too (Barnes et al., 2009; Green, Choi & Kane, 2010). Being able to forgive has been shown to predict a more benign dynamic process (Wade, Hoyt, Kidwell, & Worthington, 2014). As Wenzel et al. (2008) have argued, a general orientation towards restoration and forgiveness after victimization appears associated with more satisfactory outcomes for both victim and offender than a general orientation towards retribution and revenge. Reviews of empirical studies on the outcomes of restorative justice approaches seem to corroborate this contention (e.g., Choi, Bazemore, & Gilbert, 2012).

Knowing which kind of offenses are most likely to evoke vengeful ruminations and reactions, or, on the other hand, are more likely to be endured in silent avoidance, may help to identify individuals who are at risk of becoming long-term victims and those who are at risk of becoming offenders in turn. Empirical data regarding this question appears to be lacking; filling the gap might enhance our understanding of interpersonal motivations in response to victimization and contribute to the development of effective interventions to reduce its negative effects.

In this study we investigated whether different types of transgressions are associated with different TRIMs. To this aim, we catalogued respondents’ descriptions of the transgressions they had experienced, and related the ensuing offense types to their feelings of revenge, avoidance, and benevolence. To capture the full range of offenses that transpired in our sample (Boney-McCoy & Finkelhor, 1996) we used a descriptive rather than theory driven approach. Because research findings indicate that women and men may differ in their perception of and response to transgressions (e.g., reviewed in Miller et al., 2008; Fehr et al., 2010; also see Archer, 2009; Björkqvist, 1994), we controlled for the influence of gender.

Method

Participants

Participants were 455 adolescents (M age = 18.08 years, range from 16–26 and SD = 1.81), recruited in three educational settings: 104 undergraduate psychology students (80 women, 24 men; M age = 19.4 years, SD = 1.5), 163 students in secondary vocational education (68 women, 94 men, 1 missing; M age = 18.1 years, SD = 1.9), and 188 secondary school students (106 women, 75 men, 7 missing; M age = 17.2 years, SD = 1.5). Together these educational levels provide a fair representation of the educational range offered in our country. The undergraduate psychology students received credits for a course and the students in secondary vocational education and secondary school received a small treat for their participation.

Measures

The TRIM (McCullough et al., 1998, 2006) measures interpersonal motivation in response to a social transgression. Participants are instructed “to think of a specific person who had hurt them significantly at some time in their life” (McCullough et al., 1998, p. 1589). In the Dutch version of the TRIM (Gerlsma, Lugtmeyer, van Denderen, & de Keijser, 2016), the instruction for this incident description reads: “Many people experience injustice or violence at some time in their life. Please think of the person who has hurt you in this way. What kind(s) of injustice or violence did he/she do to you?” These incident descriptions were used to categorize the recalled incidents into different offense types. To capture the full breadth of respondents’ responses, we used both a deductive and an inductive approach in categorization (Fereday & Muir-Cochrane, 2006). A preliminary coding scheme was designed with broadly defined categories. For criminal offenses we distinguished between minor offenses, property crimes, physical violence, sexual violence, and fire-setting (Brand, 2005); for noncriminal offenses we distinguished between direct and indirect aggression (Archer & Coyne, 2005). We then used the incident descriptions provided by respondents to refine the coding scheme: Categories that were never mentioned (e.g., fire-setting, minor offenses) were omitted, whereas subcategories were added when responses reflected different behavioral expressions within a broad transgression category (e.g., the omnibus category indirect aggression was subdivided into ostracism, lack of respect, gossip and slander, violation of trust, and neglect). We only used the ultimate finer-grained coding scheme in this study, with the categories bullying (only incidents described with the word “bullying”), ostracism, gossip and slander, lack of respect (including discrimination), violation of trust (including betrayal, lying, sexual infidelity), neglect, threat (including intimidation, stalking), property (including armed robbery), physical violence (all kinds of nonsexual assault resulting in physical harm), sexual violence (all kinds of sexual assault and maltreatment), and the category declined to answer. Offense types were mutually exclusive—that is, an offense could be coded in only one offense type (e.g., “He lied to me and slept with my best friend” would be coded in violation of trust, and not also in lack of respect), and different offenses within the same offense type were coded only once (e.g., “He lied to me and slept with my best friend” would be coded as one instance of violation of trust). Note that “He lied to me and he hit me” would be coded as an instance of violation of trust for the lying and an instance of physical violence for the hitting. All offense types were coded as dichotomous dummy variables with value 1 if the incident description did, and value 0 if the description did not contain a transgression that fit the category. Codes were based on the consensus of two independent raters.

Respondents marked a timeline to indicate at what age they experienced the incident, and reported their feelings of revenge, avoidance, and benevolence towards the perpetrator in the incident description on the 18 TRIM items. The TRIM includes five items about revenge motivation (e.g., “I’m going to get even”), seven items about avoidance (e.g., “I am trying to keep as much distance between us as possible”), and six items about benevolence (e.g., “Despite what he/she did, I want us to have a positive relation again”). All items are rated on a 7-point Likert-type scale (1 = strongly disagree to 7 = strongly agree).

Factor analysis of the Dutch version of the TRIM yielded three unambiguous factors that corroborated the a priori structure (Gerlsma et al., 2016). In this study, reliability in terms of internal consistency was good (Cronbach’s α = 0.85 for revenge, α = 0.87 for avoidance, and α = 0.80 for benevolence).

Procedure

The Ethical Review Board of the University of Groningen, The Netherlands approved the study. All participants read and signed an informed consent, which described the nature and the purpose of the study. The undergraduate psychology students filled in the questionnaire digitally and online; the other participants received a questionnaire on paper during school hours, and filled in the questionnaires in the presence of a research assistant. A debriefing afterwards provided information on what to do in the case of negative thoughts or feelings as a consequence of their participation.

Results

Descriptives

Table 1 summarizes the frequencies, means, and standard deviations of scores on revenge, avoidance, benevolence motivation for different types of offences. Of the total number of 455 participants, 379 (83.3%) answered the TRIM items. The 76 missing data includes 24 (5.3%) participants who explicitly stated that they had never in their life been victimized by any kind of offense; they could therefore not fill in the TRIM items since these items refer to the perpetrator of the reported offense. The remaining 52 respondents (11.4%) are truly missing data: they did not give a description of a transgression nor did they fill in the TRIM items. These 76 respondents were omitted from the analyses because they did not provide TRIM scores. Furthermore, a number of respondents (n = 75, 16.6%) did not provide an incident description while they did fill in the TRIM items (presumably about a perpetrator whose offense they did not want to reveal). These participants were categorized as the subgroup declined to answer; their TRIM scores were included in the analyses as a separate category.

Table 1. Frequencies, means, and standard deviations of scores on revenge, avoidance, benevolence motivation for different offense types (N = 455).

			Revenge		Avoidance		Benevolence
Offense types	n	% of total sample	M	SD	M	SD	M	SD
TRIM responders	379	83.3	3.32^a, *	1.41	3.88^b, *	1.42	3.95^b, *	1.23
Bullying	75	16.5	3.07	1.39	4.18	1.60	4.16^a, *	1.04
Lack of respect	49	10.8	3.35	1.43	4.17	1.46	4.05	1.32
Violation of trust	112	24.6	3.31	1.41	4.07	1.42	4.03	1.20
Gossip and slander	79	17.4	3.33	1.49	4.18	1.39	4.09	1.06
Ostracism	38	8.4	3.09	1.60	4.08	1.66	4.33	1.27
Neglect	53	11.6^b, *	3.02	1.44	4.13	1.62	4.14	1.25
Threat	66	14.6	3.56	1.47	3.85^b, *	1.45	3.60	1.24
Physical violence	69	15.2^a, *	3.70^a, *	1.43	4.31^b, *	1.48	3.51	1.30
Sexual violence	14	3.1	3.39	1.55	5.79	0.86	4.00	0.97
Theft	38	8.4	4.16	1.40	3.98	1.57	3.35	1.30
Declined answer	75	16.6^b, *	3.54	1.44	3.56^b, *	1.04	3.76	1.15

Note. TRIM = transgression-related interpersonal motivation.

^aSignificant gender difference: women < men. ^bSignificant gender difference: women > men.

*p < .05, two-tailed.

Most respondents (n = 188, or 41.3% of the total group) reported one offense (e.g., “I was sexually assaulted”), but a substantial number of respondents (n = 115 or 25.3%) reported two or more different offenses, presumably, and as requested in the TRIM instruction, committed by the same perpetrator (e.g., “He lied to me and he hit me”). The mean number of offenses reported on the incident registration question was 1.72 (SD = 1.53, range from 0–9). The mean time that had elapsed since the incident had occurred was 4.80 years (SD = 4.08), with a range from 0–17 years.

We tested whether men and women were equally represented in all offense types. More men than women declined to describe what particular incident had happened (χ² = 10.13, df = 1, p = .001), and reported incidents involving physical violence (χ² = 15.83, df = 1, p < .001). More women than men reported incidents involving neglect (χ² = 5.87, df = 1, p < .05). Furthermore, the women reported more avoidance, t(355) = −2.88, p < .01, more benevolence, t(355) = −2.11, p < .05, and less revenge, t(355) = 3.97, p < .001. Secondary school students reported more incidents involving physical violence than the other students (χ² = 13.07, df = 2, p = .001); their mean TRIM scores did not show statistically significant differences.

Association of offense type with TRIMs

The second research question was tested with stepwise hierarchical regression analysis of TRIMs on offense type (entered in Step 2), controlled for gender (entered in Step 1), for each dependent variable. Spearman’s rho correlations were calculated to check whether TRIM scores were associated with respondents’ age, age at which the incident occurred, and the time elapsed since the incident. Only two correlations were significant (both Spearman’s rho = −0.11, p < .05, two-tailed), indicating that higher revenge scores were related to younger age and less time between the offense and TRIM measurement. To maintain comparability between outcomes for the different TRIMs and in view of the (small) effect sizes we did not include age and elapsed time in the regression analyses. Table 2 summarizes the main findings.

Table 2. Multiple correlations, R²-value change, and standardized betas for the regression of transgression-related interpersonal motivation on offense types, controlled for gender.

	Revenge				Avoidance				Benevolence
Variable	F	MR	ΔR^2	β	F	MR	ΔR^2	β	F	MR	ΔR^2	β
Gender	12.53	.21**	.05**	−.21**	4.06	.12*	.02*	.12	2.56	.10	.01	.01
Offense type:	2.20	.35**	.08*		3.56	.37**	.12**		3.32	.35**	.11**
Bullying				−.06				.11				.01
Lack of respect				.05				.00				.03
Violation of trust				.06				.05				−.04
Gossip and slander				.03				.10				.01
Ostracism				.06				−.07				.20*
Neglect				−.10				.10				−.14
Threat				.04				.09				−.14*
Physical violence				.17*				.13				−.26**
Sexual violence				−.02				.25**				.10
Theft				.22**				.05				−.18**

Note. β in the full regression equation.

*p < .05. **p < .001.

Revenge motivation

The full regression model was significant, F(11, 256) = 3.19; p < .001; the predictors together explained 13% of the variance in revenge motivation. The regression weight for gender was significant (p = .036), indicating more revenge motivation in men than in women. Offense types that, controlled for gender differences, were significantly and positively associated with increased revenge motivation were physical violence (β = 0.17, p = .019) and property crime (β = 0.22, p < .001).

Avoidance motivation

The full model was significant, F(11, 257) = 3.64; p < .001; all predictors together explained 14% of the variance. Controlled for the gender difference, which did not contribute significantly (p = .062), the only independent significant contributor to the equation was the dummy variable indicating whether or not the offense involved sexual violence (β = 0.25, p < .001).

Benevolence motivation

The regression model with all predictors in the equation was significant, F(11, 261) = 3.27; p < .001; all predictors together explained 12% of the variance. Apart from the nonsignificant contribution of gender (p = .838), results mirrored those found for revenge motivation: benevolence was lower for offenses involving property crime (β = −0.18, p = .003) and physical violence (β = −0.26, p < .001). Being the victim of threats (β = −0.14, p = .019) additionally contributed to less benevolence, whereas victims of ostracism reported more benevolence (β = 0.20, p = .046).

Discussion

We examined which types of transgressions are reported by adolescents when they are asked for their experiences with injustice and violence, and to what extent different types of offenses elicit different responses in terms of TRIMs (i.e., feelings of revenge, avoidance, and benevolence toward the perpetrator; McCullough, 2008; McCullough et al., 1998, 2006).

Respondents in this study reported a variety of transgressions, both criminal and noncriminal, and both direct as well as indirect aggression. Moreover, one in every four respondents reported more than one offense, presumably committed by the same perpetrator. These figures seem in line with epidemiologic findings on victimization in adolescents by Finkelhor et al. (2009). Those authors recommended that clinicians and researchers would need to inquire about the full range of victimization types in order to be able to identify multiply victimized children and optimally tailor interventions to all the threats that children may face. The descriptions provided by our respondents reflected salience of various subtypes of, in particular, indirect aggression, that seem in line with categories inductively inferred from previous research (Crombag et al., 2003; Rapske et al., 2010). It is interesting to note, however, that some types of victimization that are highly visible in the research literature (e.g., cyberbullying, vicarious victimization, discrimination) were hardly or not at all mentioned in this sample. This might be due to sample and context specificity of our findings, as well as perhaps an overemphasis on certain types of victimization in the research literature, as was suggested by Finkelhor et al. (2009).

Offense type did explain part of the variance in TRIMs. In particular, property crime and physical violence seemed to elicit revenge motivation and lack of benevolence. Sexual violence stood out in eliciting mainly avoidance but not revenge. Overall, these findings seem to support the notion that some types of offenses are particularly hard to forgive (Rapske et al., 2010; Van Biema et al., 1999; Worthington & Wade, 1999); moreover, it appeared to be primarily the criminal offenses that were least likely to be forgiven. Generally in line with meta-analytic findings (Miller et al., 2008; see, however, Fehr et al., 2010) and theoretical formulations (e.g., Archer, 2009; Björkqvist, 1994), we found gender differences in TRIMs, but they, too, appeared to depend on offense type. Specifically, the women in our sample reported more avoidance motivation after incidents involving physical violence, whereas the men reported more revenge after such incidents. A direct test of possible and theoretically hypothesized interactions between gender and offense types was beyond the scope of our study, but could, in future studies, contribute to our understanding of retaliatory aggression in men and women.

Implications: Research and practice

The finding that different offenses elicited somewhat different TRIMs implies that research findings on revenge and forgiveness in individuals who suffered relatively mild social transgressions (as, for instance sometimes induced in experimental settings) cannot be readily generalized to victims of criminal offenses. Hence, wherever theoretical formulations are grounded on empirical findings for noncriminal offenses, predictions made on the basis of these models should be empirically tested in victims of crime as well as victims of noncriminal transgressions.

The differences between offense types we found may be due to differences in perceived severity of the offenses: Offenses that are perceived as more severe in the community are more likely to end up in the penal code. Yet, offense severity cannot fully explain our findings. For one thing, aggressive behavior that is not liable for prosecution can nevertheless hurt, as was documented for instance in victims of bullying (e.g., Ttofi, Farrington, Lösel, & Loeber, 2011), ostracism (Saylor et al., 2013), and neglect (Welch & Bonner, 2013). Yet, these types of offenses did not appear to elicit particularly high revenge motivation. Moreover, despite neuropsychological evidence that ostracism is processed in the same way as physical pain (Williams, 2007), victims of ostracism in our study reported relatively high benevolence. Secondly, although sexual violence is doubtlessly a severe and criminal offense, victims of sexual violence appeared inclined to avoid rather than avenge the perpetrator. Hence, offense severity can only be part of the explanation, and future studies on revenge and forgiveness in response to victimization might incorporate offense type to examine whether victims of different types of offenses come to a different cost-benefit appraisal of retaliation (McCullough et al., 2013), experience different (or more intense) emotions (Aureli & Schaffner, 2013; O’Connor & Adams, 2013), and/or make different moral evaluations of the offender (Gintis, 2013). Having said that, it should be noted that the associations of offense types with TRIMs were modest, emphasizing the need to investigate individual differences such as victims’ social information processing (Dodge et al., 1990), general orientation towards justice (Wenzel et al., 2008), and implicit theories about personality (Dweck, Chiu, & Hong, 1995).

Another implication is that offense type possibly moderates associations between TRIMs and third variables. For instance, offense type may moderate the link between TRIMs and the time that has elapsed since the offense was committed. While revenge motivation is notoriously persistent (e.g., Frijda, 1994), it might dissipate more rapidly when it concerns an isolated incident of bullying as compared to chronic bullying, or as compared to an incident involving physical injuries. Such moderation effects cannot be tested in retrospect but need longitudinal research designs with repeated measurement of TRIMs.

In practice, this study suggests that the TRIM might be a practical screening instrument to assess adolescents’ ways of coping with victimization, for instance in schools, counseling, or therapy settings. It appears a suitable tool to address the nature of respondents’ victimizing experiences, it invites them to report on a wide variety of experiences with injustice and violence, including criminal offenses, and it appears to yield reliable reflections of adolescents’ feelings, thoughts, and intentions towards their aggressors. This information may help to identify individuals at risk of maladaptive coping with victimization and to signal possible needs for coaching or intervention (Worthington et al., 2011). Moreover, classrooms and schools appear to differ considerably in rates of victimization and may, perhaps unknowingly, contribute to victimization processes and outcomes for their students (e.g., Saarento, Garandeau, & Salmivalli, 2015). The TRIM could be used to monitor school climate, and as a starting point for group discussions about ways of coping with victimization, to increase empathy for and empowerment of victims. Furthermore, TRIM scores can supplement instruments for risk assessment in forensic or correctional settings (Gerlsma et al., 2016) and aid threat analysis (Meloy, Hoffmann, Guldimann, & James, 2012), for instance in the context of workplace (e.g., Kausch & Resnick, 2001) or school violence (Leary et al., 2003; Weisbrot, 2008). Within a juridical context, an indication of a victim’s TRIMs might contribute to discussions and decisions in the context of restorative justice interventions (Choi et al., 2012; Sherman & Strang, 2010). The findings in this study show that some kinds of offenses are particularly likely to elicit feelings of revenge, whereas others primarily call for avoidance of the perpetrator. In both cases, unforgiving motivations are involved that may complicate individuals’ coping with victimization (Kunst, 2011; Van Denderen, De Keijser, Gerlsma, Huisman, & Boelen, 2014).

Limitations

Because the TRIM relies on self-reported incidents, we cannot be sure that the reported incidents actually happened and fully warrant the descriptions provided. Both the description of the incident(s) and the ensuing interpersonal motivations reflect perceptions and may be subject to bias. Incident descriptions may understate or overstate the severity of the offense (e.g., labeling any unwelcome admirer a stalker) or express less unforgivingness than actually experienced (e.g., conforming to a socially desirable level).

A quarter of our respondents reported incidents that involved more than one offense. This somewhat blurs the boundaries between the different categories, and raises the interesting question whether people avenge or forgive an offender or an offense. In studies that rely on autobiographical memories of real life events, it seems ecologically more valid to address respondents’ TRIMs with regard to offenders; that is, allow respondents to describe all offenses they attribute to a particular perpetrator rather than ask them to extricate one specific offense. If TRIMs regarding one specific offense (type) are the focus of research, one could perhaps better use an experimental design.

We did not assess respondents’ perceptions of the severity of the offenses they reported. While offense types might be considered one operationalization of offense severity, subjective perceptions thereof are an important alternative (Carmody & Gordon, 2011; McCullough et al., 2013; Slotter & Finkel, 2011); combining the two approaches will shed more light on their interrelationships and respective roles as moderator/mediator in the relationship between victimization and TRIMs.

Conclusion

The question why some people respond to victimization with violent retaliation while others turn to silent avoidance or forgive and move on addresses a complex puzzle with a large number of pieces. This study shows how offense type might contribute as a small piece of that puzzle. The extent to which one tends towards revenge or forgiveness after victimization appears to depend in part on the kind of harm one has suffered: Victims of criminal and directly aggressive offenses reported less forgiving motivations than victims of noncriminal and less directly aggressive transgressions. Moreover, gender differences in responses to victimization depended on offense type, too. An important implication is that empirical findings for offenses that involve indirect aggression may not be fully generalizable to offenses that involve violence and crime.

Acknowledgments

We thank the journal’s anonymous reviewers for their very thoughtful and constructive feedback on previous versions of the paper, and gratefully acknowledge the assistance of Naomi de Ruiter, Mariëtte van Denderen, Jos de Keijser for comments on our translation of the TRIM-NL, and of Harm Aardappel, Emma Bronswijk, Carolien Gielstra, Niels Jansen, Fatma Kaptan, Kübra Keskin and Ilse Molenaar for their help with data collection.