literature review essay

(This is a sample cover image for this issue. The actual cover is not yet available at this time.)

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies are encouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

The apology mismatch: Asymmetries between victim's need for apologies and perpetrator's willingness to apologize

Joost M. Leunissen a,⁎, David De Cremer b, Christopher P. Reinders Folmer c, Marius van Dijke a

a Rotterdam School of Management, Erasmus University Rotterdam, The Netherlands b China Europe International Business School, Hongfeng Road, Pudong, China c Ghent University, Gent, Belgium

H I G H L I G H T S

► The present paper investigated the congruity between victims' and perpetrators' need for apologies ► A mismatch between victims' and perpetrators' need for apologies is observed ► This mismatch is driven by the intentionality of the transgression ► This effect was mediated by anger (victims) and guilt (perpetrators) ► This mismatch has consequences for actual apology behavior and subsequent forgiveness

a b s t r a c ta r t i c l e i n f o

Article history: Received 24 February 2012 Revised 20 November 2012 Available online 26 December 2012

Keywords: Apologies Guilt Anger Perpetrators Transgressions Forgiveness

Although previous research on apologies has shown that apologies can have many beneficial effects on victims' responses, the dyadic nature of the apology process has largely been ignored. As a consequence, very little is known about the congruence between perpetrators' willingness to apologize and victims' willingness to receive an apology. In three experimental studies we showed that victims mainly want to receive an apology after an in- tentional transgression, whereas perpetrators want to offer an apology particularly after an unintentional trans- gression. As expected, these divergent apologetic needs among victims and perpetrators were mediated by unique emotions: guilt among perpetrators and anger among victims. These results suggest that an apology serves very different goals among victims and perpetrators, thus pointing at an apology mismatch.

© 2012 Elsevier Inc. All rights reserved.

Apologizing is an effective and widely supported response to transgressions (Cohen, 1999; Darby & Schlenker, 1982; Kellerman, 2006; Meijer, 1998; Tavuchis, 1991; Van Dijke & De Cremer, 2011). From an early age, people learn to apologize when they are re- sponsible for a transgression (Schlenker & Darby, 1981). Victims of transgressions are, in turn, socialized into graciously accepting such apologies (Bennett & Dewberry, 1994; Risen & Gilovich, 2007). The pro- cess where apologies lead to reconciliation is known as the “apology– forgiveness cycle” (Shnabel & Nadler, 2008; Tavuchis, 1991).

The apology–forgiveness cycle is collectively rational because nor- mative prescriptions for perpetrators to apologize and for victims to respond with forgiveness help to preserve social relationships after

conflict. Whether these normative prescriptions actually describe an empirical reality is a question that prior research has largely failed to address. The apology–forgiveness cycle seems to assume (at least implicitly) that victim and perpetrator are both motivated to recon- cile. However, empirical studies show that victims and perpetrators often differ in their interpretations of critical aspects of transgres- sions, such as who is responsible for the transgression, its significance and its long-term effects (e.g., Baumeister, Stillwell, & Wotman, 1990; Feeney & Hill, 2006; Mikula, Athenstaedt, Heschgl, & Heimgartner, 1998). If interpretations of conflict differ so much between victim and perpetrator, then are their views on the need for apologies congruent?

In this paper, we suggest that different emotions underlie the vic- tims' and perpetrators' need for apologies: anger for the victims and guilt for the perpetrators. Since these emotions serve different functions and are activated by different types of situations, victims' and perpetra- tors' need for apologies may often be mismatched. This mismatch, we argue, can have important consequences for subsequent forgiveness and reconciliation between victim and perpetrator.

Journal of Experimental Social Psychology 49 (2013) 315–324

⁎ Corresponding author at: Rotterdam School of Management, Burgemeester Oudlaan 50, 3062 PA Rotterdam, The Netherlands.

E-mail address: [email protected] (J.M. Leunissen).

0022-1031/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jesp.2012.12.005

Contents lists available at SciVerse ScienceDirect

Journal of Experimental Social Psychology

journal homepage: www.elsevier.com/locate/jesp

Author's personal copy

Need for apologies among victims and perpetrators

An apology is generally defined as a combined statement of an ac- knowledgement of wrongdoing and an expression of guilt (Lazare, 2004; Smith, 2008; Tavuchis, 1991). Since communicating such senti- ments implies that the perpetrator believes that the transgression should not have happened and should not happen again, apologies also represent an implicit promise that the transgression will not be re- peated (Kim, Dirks, & Cooper, 2009; Smith, 2008). Apologies, therefore, imply that perpetrators distance themselves from their prior actions and admit being wrong. The effectiveness of apologies in promoting trust and forgiveness among victims has been supported by a wealth of research (see e.g., Bottom, Gibson, Daniels, & Murnighan, 2002; De Cremer & Schouten, 2008; Exline, Deshea, & Holeman, 2007; Ohbuchi, Kameda, & Agarie, 1989; Van Dijke & De Cremer, 2011).

It is important to note that apologies have rather different meanings for victims and perpetrators, and they fulfill different psychological needs. According to the needs-based model of reconciliation (Shnabel & Nadler, 2008), transgressions deprive victims and perpetrators of dif- ferent psychological needs. Victims may experience feelings of inferior- ity and anger in response to transgressions (Miller, 2001; Shnabel & Nadler, 2008). Perpetrators may suffer from fear of exclusion (Exline & Baumeister, 2000), and may therefore experience guilt (Baumeister, Stillwell, & Heatherton, 1994). Apologies provide a means for address- ing these impaired needs (De Cremer, Pillutla, & Reinders Folmer, 2010; Shnabel & Nadler, 2008). However, as victims and perpetrators require different needs to be restored, apologies serve a different func- tion for either party.

For victims, apologies represent a compensation for having been vic- timized; a symbolic compensation for the injury suffered due to the of- fense (Tavuchis, 1991), and thus apologies address the state of inequity that arises when people are transgressed against (Exline et al., 2007). Anger is an emotion that is closely linked to a need for compensation and retribution (Darley & Pittman, 2003). We therefore expect that anger, which is central to the experience of injustice and victimization (Miller, 2001), drives victims' need for apologies. To our knowledge, no research has directly tested whether anger predicts a victim's need for apologies. However, there is some indirect evidence that supports this link. Anger has been linked to reconciliation attempts (Fischer & Roseman, 2007): a negative emotional reaction towards the perpetrator still leaves the possibility for reconciliation open. Since an apology is a reconciliation tool, one would expect that victims' need for apologies is positively related to anger.

For perpetrators, apologies are means for distancing themselves from their misdeeds (Goffman, 1971), and for restoring the relationship with the victim (e.g., Bottom et al., 2002; Leunissen, De Cremer, & Reinders Folmer, 2012). We believe that guilt may play a central role in the process that makes perpetrators apologize. Perpetrators may ex- perience guilt in response to having committed an interpersonal trans- gression because such a transgression poses a threat to the relationship between the victim and perpetrator (Cryder, Springer, & Morewedge, 2012). The emotion of guilt, which is strongly related to the motivation to reconcile and improve the relationship with the victim (Baumeister et al., 1994), is likely to be central to the perpetrators' perception of the need for apologies. One would therefore expect that the guiltier the perpetrators feel, the more likely they will apologize.

In sum, apologies provide a means to fulfill the different needs of victims and perpetrators in the aftermath of transgressions. However, are the victim's and perpetrator's respective needs for apologies neces- sarily aligned with each other, as suggested by the apology–forgiveness cycle? Or in other words, are apologies provided by perpetrators when they are required by victims? We suggest that this may not be the case. Since the necessity of apologies for victims and perpe- trators is linked to different emotions, we suggest that the need for apologies may often be mismatched: apologies are given when vic- tims require them least, and not when they require them most.

This notion is best exemplified by considering the role of the inten- tionality of transgressions.

Intentionality

Intentionality refers to an individual's desires, beliefs, awareness, and abilities to perform a particular action (Malle & Knobe, 1997; Malle & Nelson, 2003). An act is regarded as intentional if the actor sets out to perform the action and succeeds. In the case of transgres- sions, this means that the actor has willfully harmed the victim. Inten- tionality is of particular interest for the present research because it is a central element in the experience of transgressions and injustice. Perceptions of intentionality influence attributions of culpability and blameworthiness for transgressions, and people's tendency to respond to them with forgiveness or retribution (Darley & Pittman, 2003; Fincham, 2000; Struthers, Eaton, Santelli, Uchiyama, & Shirvani, 2008). Importantly, intentionality has also been shown to influence the emo- tions that underlie victims' and perpetrators' apology needs, namely anger and guilt (McGraw, 1987). Therefore, intentionality may reveal when victims' and perpetrators' need for apology do or do not align.

How may intentionality affect the emotions that underlie the vic- tims' and perpetrators' need for apology, and, consequently, their per- ceptions of that need? Intentional transgressions indicate that the harm suffered by the victim was due to the perpetrator (rather than to external circumstances). Hence they evoke more feelings of injustice (Darley & Pittman, 2003; Miller, 2001) and anger than unintentional transgressions do (Berkowitz & Heimer, 1989; Betancourt & Blair, 1992; Leary, Springer, Negel, Ansell, & Evans, 1998; Quigley & Tedeschi, 1996). Indeed, the relationship between the intentionality of the transgression and anger is one of the best-established findings in the justice literature (Miller, 2001). Intentional transgressions conse- quently lead to a victim having a stronger desire for compensation and retribution (Darley & Pittman, 2003). As such, it is likely that vic- tims desire an apology particularly after intentional transgressions.

For perpetrators, the intentionality of a transgression is closely linked to guilt, being particularly experienced by perpetrators after unintentional transgressions (McGraw, 1987). According to Baumeister et al. (1994), there are two important sources of guilt. First, guilt is expe- rienced as a result of anxiety for social exclusion. After an unintentional transgression, a valuable relationship is distorted beyond the perpe- trators' will, as such, the perpetrator experiences anxiety over social ex- clusion as the victim might decide to end the relationship with the perpetrator. This anxiety results in feelings of guilt (Baumeister et al., 1994). When a perpetrator transgresses intentionally, the relationship with the victim is less likely to be important to him/her and relational deterioration is more likely to have been anticipated and considered acceptable. Thus, the perpetrator experiences less anxiety for social exclusion.

Intentionality also has important consequences for feelings of guilt because the former influences the empathy that perpetrators feel to- wards the victim. In the case of an intentional transgression, perpetra- tors are aware beforehand that they will commit the transgression (i.e., it is expected; McGraw, 1987). The perpetrator thus has had time to rationalize the transgression beforehand, thereby guarding him/ herself against feelings of guilt (Baumeister, 1999; Tsang, 2002). In con- trast, unintentional transgressions come unexpected to the perpetrator. Therefore, he/she does not have any rationalizations ready to guard him/herself against feelings of guilt. In short, these processes, anxiety for social exclusion and rationalizations, suggest that perpetrators will experience guilt particularly after unintentional transgressions and as a consequence, will want to apologize particularly after unintentional, rather than intentional transgressions.

In sum, these arguments lead us to predict a mismatch between the victims' and the perpetrators' need for apology. Because victims and perpetrators may desire apologies after different types of transgres- sions, this apology mismatch could have important consequences for

316 J.M. Leunissen et al. / Journal of Experimental Social Psychology 49 (2013) 315–324

Author's personal copy

reconciliation after different types of transgressions. Because perpetra- tors ultimately decide whether to apologize or not, it seems likely that apologies will be issued mainly after unintentional transgressions as perpetrators have the highest need to apologize after unintentional transgressions. In contrast, this mismatch would also suggest that vic- tims are unlikely to receive apologies for transgressions for which they particularly desire apologies, namely intentional transgressions. Because apologizing has been shown to have positive effects on forgive- ness (e.g., McCullough, Worthington, & Rachal, 1997; Ohbuchi et al., 1989), it stands to reason that unintentional transgressions are forgiven more often than intentional transgressions.

The present paper

The aim of the present paper is to study the incongruence between perpetrators' willingness to apologize and victims' desire to receive an apology and the subsequent effects of this incongruence on reconcilia- tion. We argue that the emotional processes that underlie the victims' and perpetrators' respective needs for apologies — that is guilt on the part of perpetrators and anger on the part of victims —may not be com- plementary, and as a consequence victims and perpetrators desire an apology at very different instances. We suggest that intentionality, which is uniquely associated with each of the above-mentioned emo- tional process, may reveal this mismatch. This incongruence in turn may have important consequences for forgiveness after the transgres- sion. We tested these predictions in three studies. Study 1 was an initial test of our ideas using an autobiographical narrative task, similar to the task designed by Baumeister et al. (1990). In study 2, we introduced another manipulation of perspective and intentionality relying on a vignette methodology. In study 3, we again relied on autobiographical narrative tasks but this time we also included measures of actual apol- ogy behavior and forgiveness after the transgression in order to explic- itly show the effects of the mismatch both on needs for apologies and behavior and subsequent forgiveness.

Study 1

Method

Participants and design In total, 202 undergraduates (97 women, M(age)=20.00, SD(age)=

1.72) participated in return for course credit. Participants were randomly assigned to a 2 (perspective: victim vs. perpetrator)×2 (intentionality: intentional vs. unintentional transgression) between-subjects design.

Procedure Participants were asked to recall an intentional or unintentional

transgression of which they were either a victim or a perpetrator. Victims were asked: Please recall a situation in which somebody else

did something (unintentionally/intentionally) to you that you experienced as unpleasant or unjust. Perpetrators were asked: Please recall a situation in which you did something (unintentionally/intentionally) that this other person experienced as unpleasant or unjust. Next, participants were asked to write a small paragraph describing the transgression. Afterwards, we assessed our manipulation check, mediating variables, and dependent variable.

Measures All questions were answered on a 1 (= not at all), to 7 (= very

much) scale.

Manipulation check We checked our intentionality manipulation in the autobiograph-

ical narratives by asking “To what extent was it the other's/your in- tention to do something unpleasant or unjust?”

Mediating variables We asked participants in the victim conditions: “How angry were you

after this other person did something unpleasant or unjust?” and partici- pants in the perpetrator conditions: “How guilty did you feel after you did something unpleasant or unjust?”

Dependent variable We assessed the need for an apology with (victim): “To what ex-

tent did you want to receive an apology from this other person?” and (perpetrators). “To what extent did you want to offer an apology to this other person?”

Results

In all the analyses of Studies 1, 2, and 3, categorical predictors were effect-coded (unintentional=−1, intentional=1; victim=−1, perpetrator=1).

Manipulation check A regression analysis with perspective and intentionality as inde-

pendent variables revealed a main effect of perspective (b=−.51, t(198)=−4.54, pb.001) and a main effect of intentionality (b=.92, t(198)=8.15, pb .001). The interaction effect was not significant (b=−.16, t(198)=−1.38, p=.17). Participants in the unintentional conditions perceived transgressions as less intentional (M=2.22, SD=1.46) than participants in the intentional conditions (M=4.05, SD=1.88). Moreover, victims (M=3.65, SD=1.93) perceived the transgression as more intentional than the perpetrators did (M= 2.65, SD=1.77).

Need for apologies A regression analysis with perspective and intentionality as inde-

pendent variables revealed a main effect of intentionality (b=−.26, t(198)=−2.09, p=.04), but not of perspective (p=.62). The main effect of intentionality showed that the need for apologies was gener- ally higher after unintentional (M=5.32, SD=1.66) than after inten- tional (M=4.81, SD=1.99) transgressions.

More importantly, this effect of intentionality was qualified by the predicted interaction between perspective and intentionality (b=− .41, t(198)=−3.25, p=.001; for cell means, see Table 1).

Planned comparisons revealed that after an unintentional trans- gression, perpetrators were more willing to apologize than victims desired an apology (b=.47, t(198)=2.63, p=.009). Conversely, when the transgression was intentional, victims desired an apology sig- nificantly more than perpetrators were willing to apologize (b=−.35, t(198)=−1.96, p=.05). Perpetrators wanted to give an apology more after an unintentional transgression than after an intentional transgres- sion (b=−.68, t(198)=−3.76, pb.001). For victims, we did not find a significant difference in the need for apologies after intentional and unintentional transgressions (b=.15, t(198)=.83, p=.41).

Mediation analyses We hypothesized that specific emotions (i.e., anger on the part of

the victim and guilt on the part of the perpetrator) would mediate the relationship between intentionality and willingness to give/receive an apology. We only measured anger among victims and guilt among

Table 1 Means (SD) for need for apologies, anger, and guilt in Study 1.

Need for apologies Anger Guilt

Victim Perpetrator Victim Perpetrator

Unintentional 4.86 (1.77) 5.81 (1.39) 5.00 (1.46) 5.81 (1.14) Intentional 5.16 (1.85) 4.46 (2.08) 5.67 (1.53) 5.04 (1.79)

Note: anger was only measured among victims; guilt was only measured among perpetrators.

317J.M. Leunissen et al. / Journal of Experimental Social Psychology 49 (2013) 315–324

Author's personal copy

perpetrators. We thus split our sample into victims and perpetrators and analyzed separately whether these specific emotions mediate the effect of intentionality on willingness to give/receive an apology. Medi- ation was tested using the PROCESS macro developed by Hayes (2012), using 5000 bootstrap resamples. The reported confidence intervals are bias-corrected bootstrap confidence intervals of the probability distri- bution of the indirect effect.

Victims. A regression analysis revealed a significant (total) effect of intentionality on anger (b=.33, t(100)=2.25, p=.03): victims were angrier after intentionally transgressions than after unintentional trans- gressions. We also obtained a significant positive effect of anger on the willingness to receive an apology (b=.52, t(100)=4.87, pb.001). Finally, the indirect effect of intentionality on the willingness to receive an apology, via anger, was significant (b=.17, S.E. = .10, 95% CI (two-sided): [.03; .42]), while the direct effect was not significant (b=−.03, S.E.=.17, t(99)=−.16, p=.87). This analysis shows that there is an indirect effect of intentionality through anger on the willing- ness to receive an apology.

Perpetrators. Our analysis obtained a significant (total) effect of inten- tionality on guilt (b=−.39, t(98)=−2.55, p=.01), meaning that perpetrators felt less guilty after intentional than after unintentional transgressions. Guilt also significantly influenced the willingness to offer an apology (b=.88, t(98)=10.31, pb.001). Moreover, the total indirect effect of intentionality on apologies through guilt was signifi- cant (b=−.32, S.E.=.13, 95% CI (two-sided): [−1.19; −.1]). The di- rect effect of intentionality on need for apologies was also significant (b=−.36, S.E.=.13, t(97)=−2.69, p=.008). This analysis thus sup- ports our prediction that perpetrators are more willing to offer an apol- ogy after an unintentional transgression than after an intentional one because they feel guiltier in the former instance than in the latter one.

Discussion

Study 1 was largely in line with our predictions. Perpetrators wanted to apologize after unintentional transgressions more than after intentional ones. This effect was mediated by guilt. Moreover, we found evidence for our proposed mismatch in the sense that perpetra- tors wanted to apologize significantly more than victims wanted to re- ceive an apology after unintentional transgressions, while perpetrators wanted to apologize significantly less than victims wanted to receive an apology after intentional transgressions. We did not find a significant difference between the intentional and unintentional conditions for victims (although the mean difference was in the predicted direction). However, the indirect effect of intentionality on victims' need for apologies, mediated by anger, was significant, showing that for victims, the need for apologies is predicted by anger.

Study 2

Study 2 was designed to extend the findings of Study 1. To experi- mentally control the type of transgression, we employed a scenario study in which participants were either the victim or the perpetrator of the same transgression. Moreover, we wanted to provide a more stringent test of the emotional processes that underlie this mismatch. While Study 1 revealed that the relationship between intentionality and need for apologies is mediated by anger (victims) and guilt (perpe- trators), we were unable to rule out that anger could also play a role in the perpetrators' willingness to apologize, and that guilt could influence a victims' desire for apologies. To show conclusively that anger medi- ates only the victims' need for apology, and that guilt mediates only for perpetrators, we measured both emotions in both the victim and perpetrator conditions in Study 2.

Method

Participants and design A total of 248 undergraduate students (126 women, M(age)=

19.68, SD(age)=1.94) participated in exchange for course credit. They were randomly assigned to a 2 (perspective: victim vs. perpetra- tor)×2 (intentionality: intentional vs. unintentional transgression) between-subjects design.

Procedure Participants were presented with a short scenario. The scenario for

the victims was (manipulation between brackets): Imagine the following situation. Your colleague (accidentally/intentionally) breaks your coffee mug by pushing it off your desk. You were very fond of this coffee mug. For the perpetrator, the scenario was: You (accidentally/intentionally) break your colleague's coffee mug by pushing it off your colleague's desk, causing it to break. Your colleague was very fond of this coffee mug. After the participants read the scenario, we assessed the manipulation checks, mediators, and the dependent measure.1

Measures All questions were answered on a 1=not at all to 7=very much

scale.

Manipulation check We checked our manipulation of intentionality with the following

item: “To what extent was it (your intention/the intention of your colleague) to break the coffee mug?”

Anger and guilt We measured anger in both perspectives by asking: “How angry

would you feel about your mug being broken?” Guilt was measured in both perspectives by asking: “How guilty would you feel about your mug being broken?”

Need for apology The need for apology was measured by asking perpetrators: “To

what extent would you want to offer an apology to your colleague?”, and victims: “To what extent would you want to receive an apology from your colleague?”

Results

Manipulation check A regression analysis with perspective and intentionality as inde-

pendent variables revealed a main effect of intentionality (b=1.57, t(244)=14.01, pb.001) and a main effect of perspective (b=−.37, t(244)=−3.26, p=.001). The interaction effect was not significant. Participants in the unintentional conditions perceived transgressions as less intentional (M=1.46, SD=.87) than participants in the inten- tional conditions (M=4.69, SD=2.25). Moreover, victims (M=3.25, SD=2.31) perceived the transgression as more intentional than per- petrators did (M=2.57, SD=2.34).

Need for apologies A regression analysis with intentionality and perspective as inde-

pendent variables revealed a main effect of intentionality (b=−.24, t(244)=−2.64, p=.009), but not of perspective (p=.42). The main effect of intentionality showed that the need for apology was

1 A potential methodological problem of the current scenario is that participant find it hard to imagine the scenario happening. We included a measure for how well the participant could imagine the scenario from happening: “How hard was it for you to imagine the described situation?” (1=not at all, 7=very much). We included this item as a control variable in our moderated multiple mediation model. Including this control variable did not significantly change the results of our analysis, the indirect effects through anger for victims and guilt for perpetrators were still significant.

318 J.M. Leunissen et al. / Journal of Experimental Social Psychology 49 (2013) 315–324

Author's personal copy

generally higher after unintentional (M=5.96, SD=1.31) than after intentional (M=5.79, SD=1.55) transgressions.

More importantly, the effect of intentionality was qualified by the predicted cross-over interaction between perspective and intention- ality (b=−.46, t(244)=−5.00, pb.001; see Table 2 for cell means).

Planned comparisons revealed that after an unintentional transgres- sion, perpetrators were more willing to apologize than victims desired an apology (b=.39, t(244)=3.00, p=.003). Conversely, when the transgression was intentional, victims desired an apology more than perpetrators were willing to apologize (b=−.53, t(244)=−4.07, pb.001). Perpetrators wanted to give an apology more after an unintentional transgression than after an intentional transgression (b=−.71, t(244)=−4.66, pb.001). Victims wanted to have an apology more after an intentional than after an unintentional transgression (b=.22, t(244)=2.07, p=.04).2,3

Anger Regression analysis with perspective and intentionality as inde-

pendent variables and anger as dependent variable revealed main ef- fects on anger of perspective (b=−.22, t(244)=−3.50, pb.001) and intentionality (b=.19, t(244)=2.97, p=.003). After a transgression, victims were angrier (M=4.11, SD=1.93) than perpetrators (M= 3.30, SD=1.83), and both were angrier after intentional transgres- sions (M=4.36, SD=2.01) than after unintentional ones (M=3.37, SD=1.72). These effects were qualified by a significant interaction ef- fect (b=−.21, t(244)=−3.34, p=.001). Simple effects analyses (Cohen, Cohen, West, & Aiken, 2003) indicated that the intentionality of the transgression significantly influenced anger among victims (b=.40, t(244)=5.52, pb.001) but not among perpetrators (b=−.02, t(244)=−.22, p=.82). Victims were angrier than perpetrators after in- tentional transgressions (b=−.43, t(244)=−4.78, pb.001), but equally angry after unintentional transgressions (b=−.001, t(244)=−.11, p=.91).

To test whether anger indeed predicts the victims' need for an apology but not the perpetrators', we conducted a regression analysis with anger and perspective as independent variables and need for an

apology as the dependent variable. We obtained a significant main ef- fect of anger (b=.65, t(244)=7.20, pb.001), but no significant main effect of perspective (b=.07, t(244)=.81, p=.42), or a significant interaction (b=.005, t(244)=.05, p=.96).

Guilt A regression analysis with perspective and intentionality as predic-

tor variables and guilt as dependent variable yielded a significant main effect of perspective (b=.42, t(244)=6.81, pb.001), indicating that after a transgression, perpetrators felt guiltier (M=4.98, SD= 1.92) than victims (M=3.19, SD=1.95) did. We did not obtain a signif- icant main effect of intentionality (b=−.10, t(244)=−1.55, p=.12), and also no significant interaction effect (b=−.10, t(244)=−1.66, p=.10). Simple effects analyses indicated that intentionality only af- fected guilt among perpetrators (b=−.20, t(244)=−1.95, p=.05), and not among victims (b=.006, t(244)=.09, p=.93). Hence, al- though the interaction term is not significant, the simple slopes analyses show a pattern on guilt consistent with our hypotheses. Nevertheless, these results should be interpreted with caution.

To test whether guilt predicts perpetrators', rather than victims' need for an apology, we conducted a regression analysis with guilt and perspective as independent variables and the need for apology as de- pendent variable. We obtained a main effect of guilt (b=.28, t(244)= 6.21, pb.001) and of perspective (b=−1.46, t(244)=−6.56, pb.001). Importantly, we also obtained the predicted interaction effect between guilt and perspective (b=.28, t(244)=6.12, pb .001). Simple effects analyses indicated that guilt only predicted perpetrators' need for apolo- gies (b=.56, t(244)=7.47, pb.001), but not the need for apologies among victims (b=.003, t(244)=.07, p=.94).

Mediation Mediation was tested using the PROCESS macro developed by

Hayes (2012), using 5000 bootstrap resamples. The reported confi- dence intervals are bias-corrected bootstrap confidence intervals of the probability distribution of the indirect effect.

We tested our model by using intentionality as the independent var- iable, anger and guilt as mediators in parallel, need for apologies as de- pendent variable and perspective as moderator, moderating the path from intentionality to anger and to guilt. In line with our hypotheses, we obtained for victims a significant indirect effect of anger (b=.24, S.E. = .06, 95% CI (two-sided): [.15; .38]) but not of guilt (b=−.001, S.E. = .02, 95% CI (two-sided): [−.03; .04]). For perpetrators, we obtained a significant indirect effect of guilt (b=−.04, S.E. = .03, 95% CI: [−.12; −.0007]), but not of anger (b=−.01, S.E. = .07, 95% CI (two-sided): [−.15; .12]). While the conditional direct (unmediated) effect of intentionality on victims' need for apologies was not significant (b=−.05, S.E. = .09, t(240)=−.50, p=.62), it was for perpetrators' apology needs (b=−.51, S.E. = .13, t(240)=−3.86, pb.001). The total effect of intentionality on the need for apologies was not signifi- cant (b=−.06, t(246)=−.93, p=.35).

Discussion

The results of Study 2 are consistent with our mismatch hypothesis. Victims have a significantly higher need for apologies than perpetrators after intentional transgressions, while perpetrators have a significantly

Table 2 Means (SD) for need for apologies, anger, and guilt in Study 2.

Need for apologies Anger Guilt

Victim Perpetrator Victim Perpetrator Victim Perpetrator

Unintentional 5.71 (1.46)

6.49 (.64) 3.38 (1.75)

3.34 (1.70) 3.17 (1.94)

5.39 (1.66)

Intentional 6.15 (1.24)

5.08 (1.85) 4.91 (1.80)

3.25 (1.97) 3.20 (1.98)

4.55 (2.10)

2 In this study, we included a measure of harm severity: (perpetrators) “To what extent would you feel that you harmed your colleague?”; (victims) “To what extent would you feel that you are harmed by your colleague?” (both are on a 1=not at all, to 7=very much scale). A regression analysis with perspective and intentionality as independent variables and harm severity as dependent variable indicated a significant main effect of perspective (b=.68, t(242)=6.62, pb.001), intentionality (b=.56, t(242)=5.45, pb.001), and a sig- nificant interaction between perspective and intentionality (b=−.27, t(242)=−2.63, p=.009). The main effect of perspective indicated that perpetrators (M=4.36, SD=1.54) considered that they harmed the victim more severely than victims felt that they were harmed (M=2.98, SD=1.71). Moreover, intentional transgressions (M= 4.11, SD=1.71) were generally perceived as more harmful than unintentional transgres- sions (M=2.80, SD=1.60). The interaction effect indicated that only victims differed in their perceptions of harm severity depending on the intentionality of the transgression: they considered intentional transgressions (M=3.84, SD=1.76) significantly (b=.82, t(242)=7.03, pb.001) more harmful than unintentional transgressions (M=2.19, SD=1.21). Perpetrators considered intentional (M=4.65, SD=1.48) and unintentional (M=4.07, SD=1.56) transgressions equally (b=.29, t(242)=1.73, p=.09) harmful. We added harm severity both as a covariate and as an extra mediator in our moderated multiple mediation model. For neither of the perspectives was the indirect effect through harm severity significant. Moreover, in both analyses, a significant indirect effect through anger and guilt remained. These analyses show that harm severity does not explain our effects.

3 An alternative explanation for why perpetrators are less willing to apologize after in- tentional than after unintentional transgressions is that perpetrators might fear that their apology will be rejected by the victim particularly after an intentional transgression. In or- der to test this alternative explanation, we measured whether fear of rejection of the apol- ogy was a concern to perpetrators with “Would you feel worried that your colleague might reject your apology in this situation?” (1=not at all, 7=completely). A regression analysis with intentionality as independent variables did not show a significant main ef- fect of intentionality. Hence, our data do not provide evidence that perpetrators were mo- re worried about an apology being rejected after intentional compared to unintentional transgressions. Moreover, inclusion of this item in as an extra mediator did not indicate a significant indirect effect through this fear of rejection item, while the indirect effect through guilt was still significant.

319J.M. Leunissen et al. / Journal of Experimental Social Psychology 49 (2013) 315–324

Author's personal copy

higher need for apologies than victims after unintentional transgres- sions. Moreover, we find that guilt only mediates the relationship be- tween intentionality and need for apologies for perpetrators, while anger mediates only the victims' need for apologies.

Two findings were not in line with our hypotheses. First we did not find a significant interaction effect between anger and perspec- tive on the need for apologies, meaning that in this study anger was predictive for the need for apologies for both victims and perpetra- tors. This might just result from testing the same effect across multi- ple studies. Even if an effect exists objectively, statistical logic dictates that some replication attempts will not show the effect (Schimmack, in press). A more substantial post-hoc explanation for this finding re- lates to the specific nature of this study. Specifically, perpetrators may have interpreted this question as being angry at themselves for the coffee mug being broken. This would be in line with our finding of a positive effect of anger on the willingness to apologize of perpetra- tors. A second finding that was not in line with our hypotheses was that, although guilt mediated the relationship between intentionality and need for apologies for perpetrators, there was still a significant direct (i.e., unmediated) effect of intentionality on the need for apol- ogies. This finding suggests that other mechanisms, besides guilt, may also play a role in the effects of intentionality on the willingness to apologize. Moral disengagement might be a likely mechanism, such as victim derogation.

Study 3

We conducted Study 3 to test whether the results of Studies 1 and 2 can be generalized to a different population (i.e., working adults). This would strengthen the relevance and scope of the mismatch between victim's and perpetrator's need for apologies. A second reason for conducting Study 3 is our aim to gain more insight into actual apology behavior and subsequent forgiveness. As explained in the introduction, apologies generally lead to forgiveness (e.g., McCullough et al., 1997). As such, we predicted that the transgression would more likely be forgiven after an apology than when no apology is given. Because apologies are more likely to be offered after unintentional than after intentional trans- gressions, this would also imply that unintentional transgressions are more likely to be forgiven than intentional transgressions.

Method

Participants and design A total of 383 working adults (286 women, M(age)=37.36,

SD(age)=10.5) were recruited through an online research participa- tion scheme of a European distance-learning university. They partici- pated for course credit. The participants were randomly assigned to a 2 (victim vs. perpetrator)×2 (intentional vs. unintentional transgres- sion) between-subjects design.

Procedure This study was conducted on the Internet and we used the same

instructions as for the autobiographical narratives in Study 1, but in this case, we asked the participants to recall a transgression from their own workplace.

Measures Unless otherwise specified, all measured were answered on a 1=

not at all, to 7=very much scale. The manipulation check and the need for apologies were measured in the same way as in Study 1. Anger and guilt were measured for both victims and perpetrators. In order to measure anger, we asked: “How angry were you after you/ this other person did something unpleasant or unjust?” To measure guilt, we asked: “How guilty did you feel after you/this other person did something unpleasant or unjust?”

Apology behavior. To measure whether an apology was issued or not after the transgression, we asked victims: “Did you receive an apology from this other person?”, and we asked perpetrators: “Did you offer an apology to the other person?” The answer scale was dichotomous: Yes or No.

Forgiveness. To check whether the transgressions were eventually for- given or not, we asked victims: “I have forgiven the other person for what he/she did.” and perpetrators: “The other has forgiven me for what I did.”

Results

Manipulation check A regression analysis with perspective and intentionality as inde-

pendent variables revealed a main effect of intentionality (b=3.27, t(379)=5.89, pb.001). The main effect of perspective was not signif- icant (p=.45). Participants in the unintentional conditions perceived transgressions as less intentional (M=1.81, SD=1.43) than partici- pants in the intentional conditions (M=3.54, SD=2.10). We also obtained an interaction between intentionality and perspective (b=−1.03, t(379)=−2.93, p=.004). This effect revealed the inten- tionality manipulation to be stronger among victims (M(intentional)= 4.35, SD=1.94; M(unintentional)=2.11, SD=1.49) than among perpe- trators (M(intentional)=2.72, SD(intentional)=1.94; M(unintentional)=1.51, SD(unintentional)=1.32). Nevertheless, both victims (b=2.24, t(379)= 9.02, pb.001) and perpetrators (b=1.21, t(379)=4.90, pb.001) rated the intentional transgressions as clearly being more intentional than the unintentional transgressions. Our hypotheses imply variations in the direction of the effect of intentionality for victims versus perpetra- tors. Hence, we do not consider these results for the manipulation check to be problematic because they indicate variations in the strength of an effect that is in the same direction for victims and perpetrators.

Content coding of the perpetration stories As an additional manipulation check for the perpetrator condi-

tions, we had all the perpetrator stories of Studies 1 and 3 (the two autobiographical narrative studies) coded by a coder blind to the original conditions and our hypotheses. An additional 20 percent was coded by a second coder to check for inter rater reliability.

The stories were coded in four categories, in line with the categoriza- tion of Darley and Pittman (2003): accidental, negligent, reckless, and intentional. In addition to this forced-choice categorization, we also had the coders rate each story on a 1 to 7 scale on the extent to which the transgression was accidental, negligent, reckless or intentional. A Chi-square analysis on the categorization of the transgression stories between the two coders showed a highly significant relationship be- tween the two coders (χ2(9)=126.23, pb .001). Correlations between the Likert scales were all high: accidental: r=.86, pb.001; intentional: r=.92, pb .001; negligent: r=.79, pb .001; reckless: r=.74, pb.001.

Of the stories, 74 were coded as accidental, 164 were coded as inten- tional, 24 were coded as negligent and 14 were coded as reckless; 16 were uncodable. These 16 cases were omitted from further analyses. This left a total of 276 cases. Of the stories written in the intentional ex- perimental conditions, 85% was coded as intentional, 1% was coded as accidental, 3% was coded as negligent and 4 % was coded as reckless. Of the stories written in the unintentional conditions, 52% was coded as accidental, 25% was coded as intentional, 15% was coded as negligent and 6% was coded as reckless. Excluding those participants whose stories were not in line with the experimental condition (e.g. described an intentional transgression in the unintentional condition), did not change the data patterns presented hereafter.

Need for apologies A regression analysis with intentionality and perspective as inde-

pendent variables revealed a main effect of intentionality (b=−.26,

320 J.M. Leunissen et al. / Journal of Experimental Social Psychology 49 (2013) 315–324

Author's personal copy

t(379)=−2.55, p=.01) and perspective (b=−.32, t(379)=−3.11, p=.002). The main effect of intentionality showed that the need for apologies was generally stronger after unintentional (M=4.52, SD=2.08) than after intentional (M=4.00, SD=2.14) transgres- sions. The main effect of perspective indicated that victims (M= 4.60, SD=2.08) generally had a stronger need for apologies than per- petrators (M=3.89, SD=2.21).

More importantly, we also obtained the predicted interaction between perspective and intentionality (b=−.58, t(379)=−5.64, pb.001; see Table 3 for cell means).

Planned comparisons revealed that victims wanted to receive an apology more after an intentional than after an unintentional transgres- sion (b=.32, t(379)=2.18, p=.03). Perpetrators wanted to give an apology more after unintentional than after intentional transgressions (b=−.84, t(379)=−5.80, pb .001). In line with the mismatch hypoth- esis, we found that perpetrators were somewhat more willing to apolo- gize than victims desired an apology after an unintentional transgression (b=.26, t(379)=1.74, p=.08). Although, the pattern in is the hypothe- sized direction, the difference is not significant and should be interpreted with caution. Conversely, when the transgression was intentional, victims desired an apology significantly more than perpetrators were willing to apologize (b=−.90, t(379)=−6.40, pb .001).

Anger A regression analysis with perspective and intentionality as inde-

pendent variables and anger as dependent variable revealed main effects on anger of perspective (b=−.56, t(379)=−13.33, pb.001) and intentionality (b=.15, t(379)=3.72, pb.001). After a transgres- sion, victims were angrier (M=5.45, SD=1.40) than perpetrators (M=3.19, SD=1.90), and both were angrier after intentional (M= 4.61, SD=2.00) than after unintentional transgressions (M=3.98, SD=1.99). These effects were qualified by a significant interaction effect (b=−.08, t(379)=−1.93, p=.05). Simple effects analyses indi- cated that victims were significantly angrier after intentional than unintentional transgressions (b=.24, t(379)=3.99, pb.001). We did not find any effect on anger among perpetrators (b=.07, t(379)= 1.27, p=.21).

To test whether anger indeed predicts the need for an apology for victims but not for perpetrators, we conducted a regression analysis with anger and perspective as independent variables and need for an apology as the dependent variable. We obtained a significant inter- action of anger and perspective (b=−.50, t(379)=−5.67, pb .001). A simple effects analysis indicated that anger only predicted the need for an apology for victims (b=.59, t(379)=5.99, pb .001), and not for perpetrators (b=−.10, t (379)=−1.44, p=.15). We also obtained a main effect of anger (b=.24, t(379)=3.97, pb.001), indicating that participants generally perceived a greater need for apologies as they became angrier.

Guilt A regression analysis with perspective and intentionality as predic-

tors and guilt as dependent variable yielded significant main effects of perspective (b=.48, t(379)=10.74, pb.001) and intentionality (b=−.11, t(379)=−2.40, p=.02). After a transgression, perpetrators felt guiltier (M=4.04, SD=1.96) than victims (M=2.15, SD=1.58), and both felt guiltier after unintentional (M=3.33, SD=2.11) than

after intentional transgressions (M=2.89, SD=1.91). These effects were qualified by a significant interaction effect between perspective and intentionality (b=−.13, t(379)=−2.99, p=.003; see Table 3 for cell means). Simple slopes analyses indicated that perpetrators felt guiltier after unintentional than after intentional transgressions (b=−.24, t(379)=−3.81, pb.001). We found no effect on guilt among victims (b=.01, t(379)=.47, p=.64).

To test whether guilt predicts the need for an apology for perpe- trators but not for victims, we conducted a regression analysis with guilt and perspective as independent variables and need for an apol- ogy as dependent variable. We obtained the predicted interaction ef- fect between guilt and perspective (b=.40, t(379)=9.15, pb.001). Simple slopes analyses indicated that guilt only predicted the need for apologies for perpetrators (b=.77, t(397)=12.47, pb.001), and not for victims (b=−.13, t(379)=−1.71, p=.09). We also obtained a main effect of guilt (b=.32, t(379)=6.49, pb.001), indicating that participants perceived a greater need for apologies as they felt guiltier.

Mediation analyses Mediation was tested using the PROCESS macro developed by

Hayes (2012), using 5000 bootstrap resamples. Like in the previous studies, the reported confidence intervals are bias-corrected boot- strap confidence intervals of the probability distribution of the indi- rect effect.

We tested our model by using intentionality as the independent var- iable, anger and guilt as mediators in parallel, need for apologies as de- pendent variable and perspective as moderator, moderating the paths from intentionality to anger and to guilt and the paths from anger to need for apologies and guilt to need for apologies. In line with our hy- potheses, for victims, we obtained a significant indirect effect of anger (b=.14, S.E. = .04, 95% CI (two-sided): [.07; .22]) but not of guilt (b=−.003, S.E. = .008, 95% CI (two-sided): [−.03; .008]). For per- petrators, we obtained a significant indirect effect of guilt (b=−.18, S.E. = .05, 95% CI (two-sided): [−.28; −.08]), but not of anger (b=−.01, S.E. = .01, 95% CI (two-sided): [−.04; .008]). The condition- al direct (unmediated) effect of intentionality on need for apologies for victims was not significant (b=−.03, S.E. = .12, t(382)=.28, p=78), while the conditional direct effect for perpetrators was significant (b=−.41, S.E. = .11, t(382)=−3.63, pb.001). The total effect of inten- tionality on the need for apologies was also significant (b=−.12, t(381)=−2.43, p=.02).

Need for apologies predicting apology behavior One of the reasons to conduct Study 3 was to investigate the be-

havioral implications of the apology mismatch. As explained in the in- troduction, because perpetrators have the highest need for apologies and unintentional transgressions and perpetrator ultimately decide whether to apologize our not, we expected that a perpetrator's need for apologies would be predictive of whether an apology was issued or not. A logistic regression analysis with perspective and need for apologies as independent variables and apology behavior as depen- dent variable indicated a main effect of need for apologies (b=1.19, Wald=56.24, pb.001) and perspective (b=.45, Wald=11.35, p=.001). We also found a significant interaction between need for apologies and perspective (b=−1.75, Wald=30.51, pb.001), showing that need for apologies was only predictive for whether an apology was issued for perpetrators (b=2.06, Wald=59.73, pb.001), but not for victims (b=.31, Wald=3.36, p=.07).

Intentionality predicting apology behavior Because perpetrators have the highest need for apologies after

unintentional transgressions, we expected that apologies are mainly is- sued after unintentional transgressions. A logistic regression analysis with perspective and intentionality as independent variables and apol- ogy issued as dependent variable yielded a main effect of intentionality

Table 3 Means (SD) for need for apologies, anger, and guilt in Study 3.

Need for apologies Anger Guilt

Victim Perpetrator Victim Perpetrator Victim Perpetrator

Unintentional 4.26 (2.04)

4.78 (2.10) 4.94 (1.56)

3.03 (1.91) 2.09 (1.45)

4.56 (1.95)

Intentional 4.89 (1.88)

3.10 (2.00) 5.89 (1.07)

3.33 (1.89) 2.20 (1.69)

3.59 (1.87)

321J.M. Leunissen et al. / Journal of Experimental Social Psychology 49 (2013) 315–324

Author's personal copy

(b=1.15, Wald=14.53, pb.001). Neither the effect of perspective nor the interaction effect was significant.

Because in logistic regression analysis lower order “main effects” are contingent upon the interaction term (Jaccard, 2001), we tested a model without the interaction term between perspective and inten- tionality. This analysis showed that compared to unintentional trans- gressions, the chance of an apology being issued after an intentional transgression becomes significantly smaller (b=1.18, Wald=29.04, pb.001, odds ratio=3.25): the likelihood of an apology being issued after an intentional transgression is significantly less than 50% (b=−.94, Wald=35.84, pb.001, odds=.39, percentage likelihood 28%). After an unintentional transgression, the likelihood of an apology being issued was equivalent to an apology not being issued at all (b=.24, Wald=2.47, p=.12, odds=1.27, percentage likelihood 56%).

Effect of apologies on forgiveness As previous research has shown that apologies aid in being forgiv-

en, we expected that perpetrators who apologized would be forgiven more than perpetrators who did not apologize. A regression analysis with apology issued (effect coded: no=−1; yes=1), perspective, and intentionality as independent variables and forgiveness as dependent variable showed a significant main effect on forgiveness of apology issued (b=.81, t(375)=8.81, pb .001), of intentionality (b=−.27, t(375)=2.93, p=.004), and of perspective (b=−.21, t(375)=−2.33, p=.02). Transgressions were generally forgiven more after an apology was issued (M=5.92, SD=1.36) than if an apology was not issued (M=4.2, SD=1.90); unintentional transgres- sions are generally forgiven more (M=5.44, SD=1.71) than inten- tional transgressions (M=4.44, SD=1.94); and victims indicated they had forgiven the perpetrator more (M=4.99, SD=1.86) than perpetrators indicated that they were forgiven (M=4.83, SD= 1.94). Neither the two-way interactions nor the three-way interac- tion were significant (p>.25).

Discussion

The results of Study 3 extend our model in a number of ways. First, we replicated our previous findings in a different population (i.e., em- ployees). Second, in line with our model, we could also show that the mismatch has consequences for actual apology behavior and subse- quent forgiveness. Whether an apology is issued or not is predicted by the perpetrator's need for apologies and not by the victim's needs. Indeed, since the perpetrator's need for apologies is higher after un- intentional transgressions than after intentional ones, apologies were issued more often after unintentional than after intentional transgres- sions. This also means that victims are unlikely to receive an apology when they have a high need for an apology and that the victim's need for an apology is not taken into account by the perpetrator when decid- ing whether to apologize or not. Finally, we were able to show that the apology mismatch has consequences for whether perpetrators are for- given or not. Perpetrators are forgiven more when they apologize. As such, unintentional transgressions are forgiven more than intentional transgressions.

General discussion

We showed across three studies that perpetrators and victims have different needs for apology, depending on the intentionality of the transgression. Victims have a stronger preference for an apology after intentional transgressions than after unintentional ones. This ef- fect is mediated by anger: victims become angrier after intentional than after unintentional transgressions, and therefore desire apolo- gies more. For perpetrators, intentionality affects the need for apology in the opposite direction: perpetrators prefer to apologize after un- intentional than after intentional transgressions, partly because they

feel guiltier after unintentional transgressions. Moreover, in Study 3 we showed that apologies are indeed issued more after unintentional than after intentional transgressions; behavior that is in line with the perpetrator's need for apologies but has no relationship to the victim's need for apologies. An apology in turn does lead to more for- giveness by the victim, as such perpetrators are forgiven more after unintentional than after intentional transgressions.

In the introduction of this paper, we argued that the apology– forgiveness cycle may not always represent an empirical reality as the victim's and perpetrator's perspectives on transgression are so di- vergent. Our findings highlight that the initiation and success of the apology–forgiveness cycle is highly dependent on the intentionality of the transgression. Perpetrators are particularly motivated to initiate the apology–forgiveness cycle by apologizing after unintentional trans- gressions. As such, unintentional transgressions are forgiven more often than intentional ones. However, in these situations (i.e., unintentional transgressions) victims are not very angry. Hence, the increased for- giveness after unintentional transgressions seems to be a joint effect of an apology and a relatively mild emotional reaction on the part of the victim. In situations where victims experience the greatest injustice and particularly desire apologies — after intentional transgressions — perpetrators are far less likely to apologize. Yet, after intentional trans- gressions, a victim's need for apologies seems to have little influence on whether an apology is issued or not. Indeed, in these situations, the ab- sence of an apology may even increase victims' anger (Ohbuchi et al., 1989). This in turn increases the risk of further escalation of the conflict. As such, intentional transgressions pose the greatest challenge for mediation and reconciliation initiatives because of the strong emotional reactions of victims combined with very incongruent reconciliatory motivations of the perpetrator.

It is interesting to note that although we find that victims generally want an apology more after intentional than after unintentional trans- gressions, related research on the effects of apologies paradoxically shows that that apologies may be of little value or even be counterpro- ductive after intentional transgressions (Struthers et al., 2008). As such, victims particularly desire an apology after intentional transgressions but at the same time, apologies seem to have limited impact in those sit- uations. What is a possible explanation for these incongruent findings regarding the need for apologies and the actual effect of apologies on victims after intentional transgressions? One potential explanation may be found in the role of forecasting errors in the apology process, whereby victims believe that they will be content if they receive an apology, but when they have actually received one, are less satisfied than they thought they would be (De Cremer et al., 2010). These find- ings again demonstrate the challenge of reconciliation after intentional transgressions: even when victims receive an apology after an inten- tional transgression (i.e., the perpetrator initiates the cycle), this may not necessarily mean that the apology is reciprocated with forgiveness.

The role of guilt in the perpetrator's willingness to apologize

Our studies showed a clear connection between feelings of guilt and the perpetrator's willingness to apologize after interpersonal transgres- sions. This is in line with recent conceptualizations of guilt, which have stressed the interpersonal effects of guilt, arguing that guilt motivates people to take relationship-restoring action (Baumeister et al., 1994; Shnabel & Nadler, 2008). Our findings connect well with this research, showing that indeed apologies as a tool for reconciliation are predicted by feelings of guilt. Focusing on the relationship between guilt and apologies therefore seems to be a promising avenue for future research on apologizing.

In this context, it is also important to distinguish guilt form other emotional reactions that perpetrators may feel after a transgression, such as compassion or sympathy. Guilt can arise when a people feel causally responsible for the harm inflicted upon the victim (Baumeister et al., 1994). As such, guilt differs from feelings of compassion or

322 J.M. Leunissen et al. / Journal of Experimental Social Psychology 49 (2013) 315–324

Author's personal copy

sympathy, which may arise when someone sees a victim suffer (i.e., from a third party perspective; Gayannée, Berthoz, Wessa, Hilton, & Martinot, 2008; Regan, 1971). Guilt only arises when people feel personally re- sponsible for the harm.

In the current set of studies, we showed that feelings of guilt have an important influence on the perpetrator's willingness to apologize. The emergence of guilt in a perpetrator is however complex. For in- stance, in this research we showed that the intentionality of the transgression is an important predictor for feelings of guilt. Some- times, however, transgressions are not easily categorized as either in- tentional or unintentional, having both intentional and unintentional characteristics. Since the premeditated nature of intentional transgres- sions provides the perpetrator with an opportunity to guard him/ herself against feelings of guilt by means of a priori rationalizations (e.g., Tsang, 2002), it seems likely that unanticipated effects of trans- gressions will make a perpetrator feel guilty. For instance, intentionally throwing a friend into the pool during a party probably does not make the perpetrator feel guilty as this was a premeditated act. However, sup- pose the friend unbeknownst had his new mobile phone in his pocket, which then broke as a result of getting wet. This unexpected effect of the transgression is likely to make the perpetrator feel guilty. Indeed, depending on the rationalizations and foreseen effects of an intentional transgression, the perpetrator may feel guilty for specific aspects of the transgression and may decide to either apologize or not.

In the present studies, we focused on guilt experienced directly after the transgression. However, when taking a longer time frame, the rela- tionship between intentionality and guilt may become more complex. Perpetrators may guard against feelings of guilt with certain ratio- nalizations. However, it seems likely that some of those rationalizations are reinterpreted later by the perpetrator and then deemed inadequate. As such, intentional transgressions may have the potential to cause guilt at a later time. Since these rationalizations are not present with unintentional transgressions, we would predict that in the long run, perpetrators may feel guiltier about intentional than unintentional transgressions, and if given the choice, would want to apologize more for something they had done intentionally than for something they had done unintentionally. It could therefore be that the apology needs of victims and perpetrators become more aligned longer after the conflict. How long this may take is of course open to empir- ical investigation.

Strengths and limitations

One of the strengths of the present research is the use of a combi- nation of different methodologies for answering our research ques- tions. We combined scenario methodology, which gives control over the transgression and thus increasing internal validity (Aronson, Wilson, & Brewer, 1998), with autobiographical narrative methodology, which is more emotionally involving and has a higher ecological validity (Baumeister et al., 1990; Gonzales, Manning, & Haugen, 1992; Zechmeister & Romero, 2002). In addition to this pluralistic methodo- logical strategy, we sampled both students and employees to test the generalizability of our results. The fact that we showed similar findings across these different methodologies and populations increases our confidence in the proposed mismatch between victims' and perpetra- tors' need for apologies.

A possible limitation of the present study is that we cannot be certain whether the task of remembering a victim episode is signifi- cantly different from remembering a perpetrator episode. Previous research comparing these perspectives also mentions this limitation (e.g., Baumeister et al., 1990). Participants might have had self- presentational concerns, selecting episodes that present themselves rather positive in their role of a considerate victim (after an un- intentional transgression) or a misunderstood perpetrator (after an in- tentional transgression). Yet, given that we find the same effects across different types of methodologies (i.e., scenario methodology and

autobiographical narrative), we feel confident that this limitation of the autobiographical narrative methodology has had no significant ef- fect on our findings.

Another important issue that must be addressed is that we only focused on a specific type of transgression, that is, anger-provoking transgressions. Victims can respond to transgressions in a number of different ways, not only with anger but also, for instance, with contempt and estrangement (Fischer & Roseman, 2007; McCullough et al., 1997). We focused on anger-provoking transgressions because anger is con- ceptualized as an emotion that can drive reconciliation (Fischer & Roseman, 2007). As such, the apology–forgiveness cycle seems to main- ly refer to anger-inducing transgressions. Yet, studying how reconcilia- tion can be achieved after contempt-inducing transgressions would be an interesting extension of the apology–forgiveness cycle. Indeed, after unintentional contempt-inducing transgressions, forgiveness may not follow as the victims are unwilling to reconcile.

On the methodological side, we relied on two different items in our analyses of our main dependent variable: one for victims and one for perpetrators. Although a direct comparison between the means on these different items (i.e., comparing perpetrator's and victim's need for apologies after intentional or unintentional trans- gressions) was important for testing our proposed mismatch, this might be problematic because these were in fact two different items. Nevertheless, by looking only at the data pattern within the victim and perpetrator conditions, it is clear that intentionality influ- ences the need for apologies of victims and perpetrators in opposite directions. Since these effects are in line with our hypotheses, we feel confident that this comparison across the different items does not pose a serious threat to the validity of our findings.

A final limitation of the current set of studies is the absence of be- havioral data after experimentally induced transgressions. Although this would be an important extension of the current findings, there are some important ethical and methodological problems with such a design. We can experimentally create unintentional and intentional transgressions with the participants as victims. However, creating sit- uations in which participants are the perpetrators presents important challenges due to the rather active role of a perpetrator compared to the passive role of a victim (Shnabel & Nadler, 2008). From a practical perspective, it seems difficult, if not impossible, to create situations in which participants intentionally transgress against one another in the lab (there are methods for creating unintentional transgressions; Leunissen et al., 2012). Moreover, creating a situation in which one intentionally transgresses against another individual might be ethi- cally undesirable as this would induce a substantial amount of stress on the research participants. Due to these considerations, we decided to test our hypotheses in scenario and autobiographical narrative methodologies only.

Concluding remarks

Due to the interpersonal nature of conflict and reconciliation be- tween the perpetrator and the victim, apologizing is a dynamic social process. Unfortunately, the psychological underpinnings of this dy- namic process have not yet been investigated in much detail. Our present results show that victims and perpetrators do not necessarily share the same perspective regarding the function of an apology, thereby making reconciliation efforts more difficult than initially anticipated.

References

Aronson, E., Wilson, T. D., & Brewer, M. B. (1998). Experimentation in social psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology, 4th ed., Vol. 1. (pp. 99–142)New York, NY: McGraw-Hill.

Baumeister, R. F. (1999). Evil: Inside human violence and cruelty. New York, N.Y.: Holt Paperback.

323J.M. Leunissen et al. / Journal of Experimental Social Psychology 49 (2013) 315–324

Author's personal copy

Baumeister, R. F., Stillwell, A. M., & Heatherton, T. F. (1994). Guilt: An interpersonal ap- proach. Psychological Bulletin, 115, 243–267.

Baumeister, R. F., Stillwell, A. M., & Wotman, S. R. (1990). Victim and perpetrator ac- counts of interpersonal conflict: Autobiographical narratives about anger. Journal of Personality and Social Psychology, 59, 994–1005.

Bennett, M., & Dewberry, C. (1994). “I've said I'm sorry, haven't I?” A study of the iden- tity implications and constraints that apologies create for their recipients. Current Psychology: Developmental, Learning, Personality, Social, 13, 10–20.

Berkowitz, L., & Heimer, K. (1989). On the construction of anger experience: Aversive events and negative priming in the formation of feelings. Advances in Experimental Social Psychology, 22, 1–37.

Betancourt, H., & Blair, I. (1992). A cognition (attribution)-emotion model of violence in conflict situations. Personality and Social Psychology Bulletin, 18, 343–350.

Bottom, W. P., Gibson, K., Daniels, S. E., & Murnighan, J. K. (2002). When talk is not cheap: Substantive penance and expression of intent in rebuilding cooperation. Organization Science, 13, 497–513.

Cohen, J. R. (1999). Apologies and organizations: Exploring an example from medical practice. Fordham Urban Law Journal, 27, 1447–1482.

Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). Mahwah, New Jersey: Lawrence Erlbaum Associates, Inc.

Cryder, C. E., Springer, S., & Morewedge, C. K. (2012). Guilty feelings, targeted actions. Personality and Social Psychology Bulletin, 38, 607–618.

Darby, B. W., & Schlenker, B. R. (1982). Children's reactions to apologies. Journal of Personality and Social Psychology, 43, 742–753.

Darley, J. M., & Pittman, T. S. (2003). The psychology of compensatory and retributive justice. Personality and Social Psychology Review, 7, 324–336.

De Cremer, D., Pillutla, M. M., & Reinders Folmer, C. P. (2010). How important is an apology to you? Forecasting errors in evaluating the value of apologies. Psychological Science, 22, 45–48.

De Cremer, D., & Schouten, B. (2008). When apologies for injustice matter: The role of respect. European Psychologist, 13, 239–247.

Exline, J. J., & Baumeister, R. F. (2000). Expressing forgiveness and repentance: Benefits and barriers. In M. E. McCullough, K. Pargament, & C. Thoresen (Eds.), Forgiveness: Theory, research, and practice (pp. 133–155). New York: Guilford.

Exline, J. J., Deshea, L., & Holeman, V. T. (2007). Is apology worth the risk? Predictors, outcomes and ways to avoid regret. Journal of Social and Clinical Psychology, 26, 479–504.

Feeney, J. A., & Hill, A. (2006). Victim–perpetrator differences in reports of hurtful events. Journal of Social and Personal Relationships, 23, 587–608.

Fincham, F. D. (2000). The kiss of the porcupines: From attributing responsibility to forgiving. Personal Relationships, 7, 1–23.

Fischer, A. H., & Roseman, I. J. (2007). Beat them or Ban them: The characteristics and social functions of anger and contempt. Journal of Personality and Social Psychology, 93, 103–115.

Gayannée, K., Berthoz, S., Wessa, M., Hilton, D., & Martinot, J. -L. (2008). An agent harms a victim: A functional magnetic resonance imaging study on specific moral emotions. Journal of Cognitive Neuroscience, 20, 1788–1798.

Goffman, E. (1971). Relations in public: Microstudies in the public order. New York: Basic Books.

Gonzales, M. H., Manning, D. J., & Haugen, J. A. (1992). Explaining our sins: Factors influencing offender accounts and anticipated victim responses. Journal of Personality and Social Psychology, 62, 958–971.

Hayes, A. F. (2012). PROCESS: A versatile computational tool for observed variable me- diation, moderation, and conditional process modeling [White paper]. Retrieved from http://www.afhayes.com/public/process2012.pdf.

Jaccard, J. (2001). Interaction effects in logistic regression. Thousand Oaks (CA): Sage. Kellerman, B. (2006). When should a leader apologize—And when not? Harvard Business

Review, 84, 72–81.

Kim, P. H., Dirks, K. T., & Cooper, C. D. (2009). The repair of trust: A dynamic bilateral perspective and multilevel conceptualization. Academy of Management Review, 34, 401–422.

Lazare, A. (2004). On apology. New York: Oxford University Press. Leary, M. R., Springer, C., Negel, L., Ansell, E., & Evans, K. (1998). The causes, phenome-

nology, and consequences of hurt feelings. Journal of Personality and Social Psychology, 74, 1225–1237.

Leunissen, J. M., De Cremer, D., & Reinders Folmer, C. (2012). An instrumental perspec- tive on apologizing in bargaining: The importance of forgiveness to apologize. Journal of Economic Psychology, 33, 215–222.

Malle, B. F., & Knobe, J. (1997). The folk concept of intentionality. Journal of Experimen- tal Social Psychology, 33, 101–121.

Malle, B. F., & Nelson, S. E. (2003). Judging mens rea: The tension between folk con- cepts and legal concepts of intentionality. Behavioral Sciences & the Law, 21, 563–580.

McCullough, M. E., Worthington, E. L., & Rachal, K. C. (1997). Interpersonal forgiving in close relationships. Journal of Personality and Social Psychology, 73, 321–336.

McGraw, K. M. (1987). Guilt following transgression: An attribution of responsibility approach. Journal of Personality and Social Psychology, 53, 247–256.

Meijer, A. J. (1998). Apologies: What do we know? International Journal of Applied Linguistics, 8, 215–231.

Mikula, G., Athenstaedt, U., Heschgl, S., & Heimgartner, A. (1998). Does it only depend on the point of view? Perspective-related differences in justice evaluations of neg- ative incidents in personal relationships. European Journal of Social Psychology, 28, 931–962.

Miller, D. T. (2001). Disrespect and the experience of injustice. Annual Review of Psychology, 52, 527–553.

Ohbuchi, K., Kameda, M., & Agarie, N. (1989). Apology as aggression control: Its role in mediating appraisal of and response to harm. Journal of Personality and Social Psychology, 56, 219–227.

Quigley, B. M., & Tedeschi, J. T. (1996). Mediating effects of blame on feelings of anger. Personality and Social Psychology Bulletin, 22, 1280–1288.

Regan, J. W. (1971). Guilt, perceived injustice, and altruistic behavior. Journal of Personality and Social Psychology, 18, 124–132.

Risen, J. L., & Gilovich, T. (2007). Target and observer differences in the acceptance of questionable apologies. Journal of Personality and Social Psychology, 92, 418–433.

Schimmack, U. (in press). The ironic effect of significant results on the credibility of multiple-study articles. Psychological Methods.

Schlenker, B. R., & Darby, B. W. (1981). The use of apologies in social predicaments. Social Psychology Quarterly, 44, 271–278.

Shnabel, N., & Nadler, A. (2008). A needs-based model of reconciliation: Satisfying the differential emotional needs of victims and perpetrators as a key to promoting rec- onciliation. Journal of Personality and Social Psychology, 94, 116–132.

Smith, N. (2008). I was wrong: The meaning of apologies. Cambridge, NY: Cambridge University Press.

Struthers, C. W., Eaton, J., Santelli, A. G., Uchiyama, M., & Shirvani, N. (2008). The effects of attributions of intent and apology on forgiveness: When saying sorry may not help the story. Journal of Experimental Social Psychology, 44, 983–992.

Tavuchis, N. (1991). Mea Culpa: A sociology of apology and reconciliation. Stanford: Stanford University Press.

Tsang, J. -A. (2002). Moral rationalization and the integration of situational factors and psychological processes in immoral behavior. Review of General Psychology, 6, 25–50.

Van Dijke, M., & De Cremer, D. (2011). When social accounts promote acceptance of unfair ultimatum offers: The role of the victim's stress responses to uncertainty and power position. Journal of Economic Psychology, 32, 468–479.

Zechmeister, J. S., & Romero, C. (2002). Victim and offender accounts of interpersonal conflict: Autobiographical narratives of forgiveness and unforgiveness. Journal of Personality and Social Psychology, 82, 675–686.

324 J.M. Leunissen et al. / Journal of Experimental Social Psychology 49 (2013) 315–324