Apologies with explicit acknowledgements of wrongdoing are especially likely to be well received. An implication of Trope's (1986) dual process model of social information processing is that this may not be so when interpersonal transgressions are ambiguous. In such cases, full apologies primarily serve to identify behaviors as affronts. In Study 1, participants read vignettes involving clear or ambiguous transgressions for which 1 of 3 alternative forms of apology were received by the wronged person: no apology, a full apology, or an expression of sympathy offered as partial apology. For clear transgressions, participants expected full apologies to soothe the wronged person more than either partial apologies or no apologies, but when transgressions were ambiguous, full apologies were considered to be less likely to ameliorate the wronged person's bad feelings than were partial apologies. In Study 2 we replicated the effect for ambiguous transgressions even when participants had the vignette presented to them as if they were in the role of the wronged person, and the results of Study 3 indicate that these findings are not an artifact of differences in conclusions about whether a transgression had actually taken place. Reactions to apologies are an interactive effect of the nature of the apology and the nature of the transgression.
Keywords: apology, insult, social inference, interpersonal transgression.
In 2009, Lloyd Blankfein, the Chief Executive Officer of Goldman Sachs, publicly apologized for his company's role in the previous year's economic meltdown by acknowledging that company personnel had "participated in things that were clearly wrong and have reason to regret". As at least one observer noted, though, "Apologies are always tricky business ... people who might otherwise have held Goldman blameless for the financial crisis now have evidence that maybe the bank does deserve some blame. If the CEO is apologizing, it must be because the company did something wrong" (Walsh, 2009). Indeed, the speech did not do much to pacify Goldman Sachs's critics.
The reaction to Blankfein's apology represents a more general phenomenon. For example, an Australian blogger speculating about the reluctance of his country's Prime Minister to apologize to Aboriginal people for their past mistreatment suggested that "the main reason that he wouldn't publicly apologize on behalf of our country was that he was afraid of the backlash. He feared an apology would mean admitting guilt and that this would fuel the disturbance rather than remedy it" (O'Leary, 2008).
Apologies, even sincere ones, do not always have their intended effects. In the research reported here we test the hypothesis that this might be especially true when--from the perspective of the recipient of the apology--there had been some ambiguity as to whether an interpersonal transgression had actually taken place.
When do Apologies Make People Feel Better--and When do They Not?
After being insulted or injured by another person, why would simply hearing the words "I'm sorry" prompt forgiveness? Apologies make people feel better for a number of reasons. First of all, an apology can elicit empathy for the perpetrator, especially when the offending person includes an explanation of the thoughts and feelings that led to the transgression (McCullough, Worthington, & Rachal, 1997). Second, by suggesting new attributions for the bad behavior (e.g., "due to being under extreme pressure" vs. "due to being a nasty person") an apology can help reassure others that the behavior will not reoccur (Gold & Weiner, 2000). Apologies also help victims restore "face" and validate their perceptions that they have suffered an injustice (Eaton, Struthers, & Santelli, 2006; Goffman, 1971); thus, an apology can serve to reduce a perceived injustice gap (Worthington, 2003).
Finally, apologies can signal commitment to a relationship and concern for the other person's feelings (Exline, Deshea, & Holeman, 2007; Lazare, 2004). The more respect people have for others and the more ongoing relationships with others are valued, the more likely people are to offer apologies to those whom they have insulted or injured. An apology can serve this function even without an explicit admission of guilt.
However, apologies may be seen as insufficient if the harm suffered by the wronged person is felt to be extreme (Darby & Schlenker, 1982). Apologies will also fail to have their intended effects when recipients believe them to be insincere or manipulative (Risen & Gilovich, 2007; Skarlicki, Folger, & Gee, 2004).
The Role of Ambiguity
Apologies often include an admission of guilt. When it is clear that a transgression was committed, such confessions do not tell recipients anything that they did not already know. Nonetheless, apologies that explicitly acknowledge wrongdoing have been found to be more effective than those that do not (see e.g., Scher & Darley, 1997).
The situation would seem to be different, though, when there is uncertainty about whether or not an act was committed for which an apology is necessary. For example, saying "I'm sorry I insulted you", confirms that the individual did, in fact, make an unwarranted nasty remark. When there is ambiguity as to whether or not any transgression was committed in the first place, an apology is informative. As a result, when an individual offers a full apology of this kind to another person, what that person might be hearing is not "I'm sorry", but instead, "I did you wrong".
The main hypothesis tested in the current research was that participants in our study would not expect that making full apologies, which involve accepting blame and acknowledging responsibility (that is to say, repentance--see Eaton et al., 2006), would elicit positive reactions from the recipients of the apologies when the related transgressions were ambiguous. Thus, we predicted that apologies may be poorly received even when they are sincere and even when the wrongdoings for which they are meant to atone are not extremely egregious.
This analysis is consistent with, and could be derived from, Trope's (1986) dual process stage model of social information processing. Although primarily focused on dispositional attribution, this model can be applied to a broader range of judgments. According to Trope, information used to interpret and react to a behavior--that is, situational information, nonverbal cues, information about a person's past behavior, verbal utterances, and so on--is used in two ways that correspond to two steps. Such information can play a role in inferential or attributional processes. For example, learning that someone with a calm expression was in a frightening situation, such as watching a gruesome horror movie, would lead an observer to conclude that this individual was calmer and more controlled than someone else in that same situation who had a panicked expression. Similarly, learning that someone apologized after making an insulting comment would lead an observer to conclude that the transgressor was less nasty, and more worthy of forgiveness, than someone who had made an insulting comment and did not apologize.
But Trope (1986) suggests in his model that the same information or input also helps determine the output of an earlier identification/categorization stage. That is to say, before thinking systematically about the behavior of another person, the individual must first determine what that behavior is. Knowing that someone is in a frightening situation makes it more likely that his or her facial expression will be interpreted as being "frightened". Similarly, when an individual learns that someone apologized after making an ambiguously rude comment, this knowledge makes it more likely that the individual will interpret the comment as having been inappropriate or insulting.
The impact that a given cue has at the different stages in the processing of behavioral information is a function of the ambiguity of the behavior being interpreted (Trope, 1986; Trope, Cohen, & Alfieri, 1991). When the behavior is ambiguous, so that the observer's question is: "What does that expression/ comment mean?", the effects at the identification stage will be greater. When it is unambiguous--when, for example, a facial expression is definitely a grimace and a comment is clearly an insult--the only measurable effects will be at the later attributional stage. Thus, when transgressions are ambiguous, full apologies will serve to identify those transgressions as having been insults or injuries; when transgressions are unambiguous, a full apology will mostly serve to reduce negative feelings.
Two Kinds of Apologies
The preceding analysis implies that a certain kind of apology would not necessarily backfire when there is ambiguity about whether one had engaged in behavior that might call for an apology. Although from some theoretical perspectives, for an interpersonal communication to be called an "apology", it must, by definition, involve an acceptance of responsibility or an admission of guilt, in this study we adopted Robbennolt's (2003, 2006, 2008) terminology by distinguishing between full apologies and partial apologies. A partial apology is "a statement that expresses sympathy, but does not admit responsibility" (Robbennolt, 2003, p. 469) and is similar to Fehr and Gelfand's (2010) expression of empathy. People sometimes simply express regret that something unpleasant and upsetting occurred; in short, apologies sometimes boil down to "I'm sorry about what happened and I'm sorry you feel badly". As such apologies are not admissions of guilt, they would not directly imply that an ambiguous incident should be identified or categorized as a wound inflicted by the apologizer. They would, however, serve one of the functions of apologies already noted: a signal of concern about other people's feelings and of the value that the apologizer placed on his or her relationship with those people. Thus, a partial apology of this sort could be expected to reduce a recipient's level of negative affect. On the other hand, for clear transgressions, partial apologies without admissions of guilt might seem less than satisfactory (Lazare, 2004). Indeed, Robbennolt (2003, 2006) found exactly that: People who are certain that they have been wronged perceive such expressions of sympathy as being manipulative or simply not sufficient to close the injustice gap.
In sum, we predicted that (a) when interpersonal transgressions were clear and unambiguous, participants would expect that full apologies with confessions of guilt would make the wronged person feel better than either partial apologies or no apologies, and that the reactions of people who had been wronged to partial apology or no apology would not differ from each other, and (b) when interpersonal transgressions were uncertain and ambiguous, participants would expect that partial apologies would make the wronged person feel better than either full apologies or no apologies, and that the reactions of the wronged person to full or no apology would not differ from each other.
Participants. Five hundred and one students (264 female and 237 male) at Syracuse University participated in the study. Participants were volunteers recruited by other students enrolled in a social psychology class taught by the first author.
Materials and procedure. Participants read a single vignette involving either a clear or ambiguous interpersonal transgression followed by either a full apology, a partial apology, or no apology (see Appendix A1 for an example). Four different basic vignettes were developed. One concerned a student who was (or might have been) caught by his girlfriend flirting with another woman; another was about a young man whose lunch was (or might have been) removed from a refrigerator and eaten by a roommate; another told a story about a college student with a secret that was (or might have been) revealed to other people by one of her friends; and one was about a woman who insulted (or might have insulted) her boyfriend's aunt.
Full apologies always included an admission of guilt and an expression of sorrow. In the case of partial apologies, transgressors did not admit to being aware of having done anything wrong, but said that they were "feeling bad about what happened". In the no apology condition, transgressors did not admit guilt, nor did they offer an apology of any kind.
After reading one of the versions of one of these vignettes, participants used a 7-point scale (ranging from much less to much more) to indicate what they thought the injured party's immediate reaction would be--specifically, whether he or she would "get more angry and upset or less angry and upset". Finally, participants reported their gender.
Results and Discussion
Data were analyzed using a 2 (Ambiguity of transgression: clear or ambiguous) x 3 (Apology type: none, partial, full) x 2 (Gender: male, female) analysis of variance (ANOVA). The ANOVA revealed expected main effects of ambiguity, F(1, 489) = 24.56, p < .001, [[eta].sup.2] = .048, and apology type, F(2, 489) = 8.62, p < .001, [[eta].sup.2] = .034. Compared with ambiguous transgressions, participants believed that clear transgressions would cause people to become more upset (5.09 vs. 4.43). In addition, participants believed that people would be more upset by a lack of an apology (M = 5.20) than by any sort of apology (partial: M = 4.50; full: M = 4.57). Post hoc least significant difference (LSD) analyses revealed that the differences between the mean for the no apology condition and the other two conditions were both significant (p < .001), while the means for the latter two conditions did not differ from each other.
As expected, the main effects were qualified by a significant Ambiguity x Apology type interaction, F(2, 489) = 14.54, p < .001, [[eta].sup.2] = .056. When a clear transgression had taken place, participants expected the wronged person to be less upset when a full apology was received than when either a partial apology or no apology was received (both p < .001; see Figure 1). When a partial apology was offered by the transgressor, participants expected the wronged person to be no less upset than he or she would have been in the absence of an apology.
When the transgression was ambiguous, however, participants believed that a partial apology would make people feel better than if they had received no apology at all, and that a partial apology would make them feel even better than if they had received a full apology (both p < .001; see Figure 1). In sum, when characters in our vignettes were described as having clearly engaged in offensive or thoughtless behavior, our participants considered a partial apology to be no better than no apology at all, but when there was some doubt as to the meaning of the actions of the characters in the vignettes, it was a full apology that was considered to be no better than no apology at all.
The ANOVA also revealed a main effect of gender, F(1, 489) = 4.72, p < .05, [[eta].sup.2] = .010. Women judged the wronged people in the scenarios to be more upset than men did (4.90 vs. 4.59). This gender difference is consistent with Schumann and Ross's (2010) finding that women have a lower threshold for what constitutes offensive behavior than do men. But the results of the analysis in our study also could have been a consequence of the specific contents of the scenarios, and thus are not readily interpretable. Gender of participant did not interact with any of the other variables. An additional analysis incorporating the vignette as a variable revealed a main effect of vignette, F(1, 478) = 13.13, p < .001. Overall, participants believed that the student who caught her boyfriend flirting and the one whose secret had been revealed to be most upset (5.14 and 5.07, respectively), and that the young man whose girlfriend insulted a relative and the one whose sandwich was stolen would be less upset (4.24 and 4.56, respectively). The analysis also revealed an Ambiguity x Apology type x Vignette interaction, F(1, 478) = 3.21, p < .005. However, the pattern was difficult to interpret; given the large sample size, the study might have been overpowered to detect this effect, which was associated with an eta squared of only .039.
The direction of the condition means for clear transgressions in Study 1 was as predicted, and also replicated existing findings (Robbennolt, 2003, 2006). More novel were the findings for ambiguous transgressions: participants believed that people would be less upset by partial apologies than by full ones. Study 2 was run to replicate those findings, but also to examine whether they might have been an artifact of the detached, third-person perspective of participants in Study
1. People often expect others' emotional reactions to be weaker than their own would be in the same situation (McFarland & Miller, 1990). It follows that people imagining themselves in the situations described in the vignettes might not be as inclined or willing to report being satisfied by partial apologies. Thus, in Study
2, half of the participants were asked to imagine themselves to be the victims of the interpersonal transgressions.
Participants. Two hundred and eight students at Syracuse University participated in the study in partial fulfillment of a course requirement. Four participants aged between 39 and 57 years were dropped from analyses as the vignettes were developed in consultation with young adult college students and featured young adults as protagonists. The final sample consisted of 108 female and 96 male participants.
Materials and procedure. The materials and procedure were similar to those of Study 1 with the following exceptions:
Unlike in Study 1, each participant was presented with two ambiguous transgressions. Half of the participants read vignettes about fictional people and gave their opinion on what they considered would be the protagonists' likely responses. For the other half of the participants, the vignettes were written in such a way that the participant was described in the vignette as the protagonist (i.e., as "you", being the offended party) so that participants were encouraged to take the protagonist's perspective. They were then asked "What would your immediate reaction be?" (see Appendix A2 for an example).
Given that half of the participants in Study 2 were being asked to take the perspective of the protagonists, male participants were presented with only the two vignettes in which the offended party was male (the lunch and insult stories). Similarly, female participants were presented with only the two vignettes in which the offended party was female (the flirting and secret stories). The order in which the two stories were presented was randomized.
Results and Discussion
Responses to the two vignettes were averaged, and data were analyzed with a 3 (Apology type: none, partial, full) x 2 (Perspective: own vs. third person) x 2 (Gender: male or female) ANOVA. Similar to Study 1, the ANOVA revealed a main effect of gender, F(1, 192) = 71.69, p < .001, [[eta].sup.2] = .272. Women believed that the wronged people in the scenario would be more upset than men did (5.82 vs. 4.63). As in Study 1, this difference is difficult to interpret with any confidence--even more so in this case, as the male and female participants were presented with different vignettes.
In addition, the expected main effect of apology type was also significant, F(2, 192) = 3.35, p < .05, [[eta].sup.2] = .034. As in Study 1 (ambiguous condition), participants believed that people (including themselves) would feel least upset in response to a partial apology (M = 5.06) as compared to either no apology (M = 5.20) or a full one (M = 5.50).
Post hoc (LSD) analyses revealed that the difference between the means for the partial and full apology conditions was significant (p < .05). The difference between the partial and no apology conditions, although not statistically significant, was in the same direction. The difference between the no apology and full apology conditions also did not reach significance, but there was a suggestion that participants might sometimes feel more angry and upset in response to a full apology than when no apology at all was offered. In sum, when there was doubt as to the meaning of perpetrators' actions, a partial apology again seemed more likely to elicit a favorable response than a full one.
There were, however, no significant main or interactive effects involving perspective (all p > .30).
Study 3 was run to address two further limitations of the previous studies. In both Studies 1 and 2, the dependent variable was a rating of how angry and upset a wronged person might be. In addition to being a single item, this is also, arguably, a double-barreled question; the person could be upset but not necessarily angry, because anger is a more specific emotional state. In Study 3, participants simply rated how angry they would expect to be. In addition, they also reported how much they expected that sadness, another basic emotion, would be triggered by the transgression.
Of even more importance for interpreting the findings in Studies 1 and 2 was controlling for an alternative account for the apparent effectiveness of partial apologies following ambiguous interpersonal transgressions. Participants might have interpreted the lack of an admission of guilt as evidence that the person who was trying to smooth things over had not, in fact, committed the transgression. If so, the dampening effect of partial apologies on negative reactions might not necessarily have been a function of inferring that the other person was signaling commitment to an ongoing relationship and concern for others' feelings. Instead, it could have been a function of concluding that an infraction had not, in fact, taken place. If so, then the findings in the previous studies would be replicated only when participants interpreted the partial apology as a denial of wrongdoing and when they accepted that denial. This possibility was investigated in Study 3.
Participants. One hundred and eighteen students at Syracuse University participated in the study in partial fulfillment of a course requirement. Three older participants were again dropped from analyses; the final sample consisted of 69 female and 46 male participants.
Materials and procedure. The materials and procedure were almost identical to those of Study 2. Again, participants responded to two vignettes. However, in Study 3, all participants were asked to take the protagonist's perspective and the single-item measure used in Studies 1 and 2, "Would you get more angry and upset or less angry and upset?", was replaced with two questions: "How angry would you be?" and "How sad would you be?". As in the previous two studies, participants were asked to report what they thought their immediate reaction would be using 7-point scales (ranging from not very to very). The order of the two questions was randomized among participants.
Finally, the last question asked after each vignette was whether or not the participant believed that the transgression had taken place (e.g., "Do you believe that your roommate ate your sandwich?"--"yes" or "no").
Results and Discussion
Overall analyses. Responses to the two vignettes were again averaged, and ratings of anger and sadness were analyzed with a 3 (Apology type: none, partial, full) x 2 (Gender: male or female) ANOVA. Consistent with the findings in Study 2, the main effect of apology type on anger was significant, F(2, 114) = 4.09, p < .05, [[eta].sup.2] = .070. Participants believed that they would feel angrier in response to a full apology (M = 5.60) than to either a partial apology (M = 5.10) or no apology at all (M = 5.00). Post hoc (LSD) analyses revealed that the difference between the means for the partial apology and full apology conditions was significant (p < .05), as was the difference between the means for the no apology and full apology conditions (p < .01).
As in Studies 1 and 2, the ANOVA also revealed a main effect of gender, F(1, 109) = 26.74, p < .001, [[eta].sup.2] = .197. The mean level of anger reported by women was higher than the mean level reported by men (5.60 vs. 4.65). The ANOVA also revealed an Apology type x Gender interaction, F(1, 114) = 6.45, p < .005, [[eta].sup.2] = .106. However, this interaction was not anticipated, not obviously meaningful, and not in evidence for any other dependent variable across the three studies.
A similar analysis of sadness revealed only a gender main effect, F(1, 109) = 22.39, p < .001. The mean level of sadness reported was higher for women than for men (4.46 vs. 3.23). However, the vignettes used for the current research would, arguably, describe situations that, from an appraisal theory perspective, would trigger more anger than sadness (Mikula, Scherer, & Athenstaedt, 1998; Roseman & Smith, 2001). The transgressions described in the vignettes presented to participants in these studies all lent themselves to an active response --specifically, confronting the offending person. Indeed, mean anger ratings were significantly higher than sadness ratings overall (5.22 vs. 3.97, t(114) = 8.51, p < .001). For situations involving hurtful behaviors such as a close other disappearing for an extended period of time, to which there is no obvious effective response, apology type could well account for more variance in sadness than anger.
Analyses restricted to those who believed a transgression occurred.
We were not surprised that in the full apology condition participants almost universally believed that a transgression had taken place (97.2% of all responses). We were also not surprised with the findings that participants were less certain about the occurrence of a transgression in the no apology condition (61.2% of all responses) and the partial apology condition (69.2%).
However, an analysis of the expected anger data restricted to judgments associated with transgressions that participants were certain had occurred yielded results similar to those found for the full set of data. For this analysis, if a participant answered "yes" twice when asked whether rudeness, betrayal, theft, or infidelity had, indeed, taken place, as in the overall analysis, the two anger judgments were averaged and used for the analysis. Alternatively, if a participant answered "yes" just once, only the anger judgment associated with that response was used for the analysis. Finally, four participants in the no apology condition and five participants in the partial apology condition were dropped from the analyses because they answered "no" twice.
The 3 (Apology type: none, partial, full) x 2 (Gender: male or female) ANOVA again revealed main effects for gender, F(1, 100) = 27.94, p < .001, [[eta].sup.2] = .218, and the predicted main effect of apology type, F(2, 100) = 3.12, p < .05, [[eta].sup.2] = .059. Participants expected that they would feel angrier in response to a full apology (M = 5.65) than in response to either a partial apology (M = 5.26) or no apology at all (M = 5.47). Due to the diminished statistical power of this analysis, post hoc comparisons between cell means did not reveal differences at p < .05. However, this analysis provides evidence that the significant effects of apology type found in all three studies were not simply an artifact of confounds between apology type and conclusions about the perpetrators' guilt.
In past research it has been found that apologies with explicit acknowledgements of wrongdoing are especially likely to be well received. In Study 1 our results replicated that finding for clear interpersonal transgressions. But in Studies 1, 2, and 3 our results revealed a limitation on that effect: when there is ambiguity as to whether an interpersonal transgression actually took place, such an apology might be heard and function primarily as a confession. In such cases, an apology might be received no more graciously--and might sometimes be received less graciously--than a lack of any apology at all. This finding is consistent with the implications of Trope's (1986) dual process stage model of social information processing.
On the other hand, even when transgressions are ambiguous, conciliatory words might help ameliorate the wronged person's distress. We found that partial apologies consisting simply of an expression of concern for the other person's feelings were seen as being more likely than full apologies to reduce anger and other negative feelings. In contrast, partial apologies were not looked on as being significantly more effective than no apology at all when transgressions were clear and unambiguous. Overall, then, immediate reactions to apologies seem to be affected interactively by the nature of the apology and the nature of the transgression.
The pattern of means was not identical across studies. In all three studies, when the transgression was ambiguous, participants considered it likely that full apologies would lead to significantly more negative affect than would partial apologies. In comparison to the two other conditions, reactions to the absence of any apology were more inconsistent. As was revealed in Study 3, this might reflect the fact that participants in this condition were most likely to be uncertain about whether an interpersonal offense had actually occurred. However, in none of the studies did our participants believe that full apologies would elicit reactions that were significantly more positive than no apologies at all when the transgression was ambiguous, and in Study 3 participants thought that a full apology would trigger significantly more anger.
The findings reported here apply to immediate reactions, and, thus, do not exclude the possibility that conversations following a full apology could eventually lead to reconciliation and forgiveness, even when the wrongdoing was ambiguous. In addition, the severity of the wrongdoing (Darby & Schlenker, 1982) might moderate reactions to apologies even in the case of ambiguous transgressions, limiting the extent to which these findings can be generalized to other cases of wrongdoing. This might explain why Robbennolt (2003, 2006, 2008), who manipulated ambiguity in some of her experiments, reported a more complex and inconsistent pattern of results than we have reported here. In her vignettes she described an automobile accident in which a bicyclist was injured, which is an event with more negative consequences than those described in our vignettes. The fact that Robbennolt's (2003, 2006, 2008) vignettes were presented to participants as evidence used in legal proceedings might also have played a role in participants' judgments, as could the manipulation of other variables relevant to that particular context, such as the rules of evidence. In addition, participants in Robbennolt's (2008) study were practicing attorneys.
Nor should our findings be taken to imply that people should not apologize when their behavior is ambiguously hurtful to others. That is a moral issue, not an empirical one. The current findings simply shed light on why sometimes an apology is not immediately met with the gratitude and relief that the apologizer might have expected.
A limitation in this research is the lack of direct assessment of mediating variables. Future researchers could more directly assess whether partial apologies are well received after an ambiguous transgression because they signal commitment to a relationship. Similarly, researchers could more directly assess whether apologies of that type are ineffective in the case of clear interpersonal wrongdoing because they are seen to be manipulative or evasive or because they are inadequate for restoring justice and equity.
Finally, both the reliance on vignettes in this research and the focus on anticipated rather than actual reactions limit our ability to generalize these findings to real-life transgressions. De Cremer, Pillutla, and Folmer (2011) found that people often overestimate how much they would value receiving an apology. It is, however, significant that participants differentiated between anticipated anger and sadness in Study 3. This suggests that they were not simply noting whether outcomes were pleasant or unpleasant or responding in a similarly thoughtless, superficial manner, but were trying to simulate the psychological situations described by the vignettes.
We doubt that knowledge of any of our findings could have helped Goldman Sachs rehabilitate its public image. But the research reported here indicates that relationships between characteristics of apologies and their effectiveness are neither simple nor linear.
The full text of all vignettes is available from the first author on request.
A1. Sample vignette
Colleen had an embarrassing experience with a guy from one of her classes--in brief, she thought that he was interested in her, but she was quite wrong about that. She told two of her close friends (Jen and Allison) about what had happened, and asked them to keep it a secret.
AMBIGUOUS: Two days later, though, she glanced at the screen of an acquaintance's laptop, and saw that the acquaintance was reading an email message about her embarrassing experience. She couldn't be sure, but she thought she saw something that suggested that the message was from Jen. All day long she wondered about whether Jen had revealed her secret.
FULL APOLOGY: Later, Colleen ran into Jen and told her what she had seen. Jen admitted to sending the message and said she was sorry.
PARTIAL APOLOGY: Later, Colleen ran into Jen and told her what she had seen. Jen said that she really couldn't recall having talked about Colleen's embarrassing experience with anyone but Allison, but told Colleen that she was sorry that Colleen was feeling bad about what happened.
NO APOLOGY: Later, Colleen ran into Jen and told her what she had seen. Jen did not admit to having done anything and did not offer any apologies.
CLEAR: Two days later, though, she glanced at the screen of an acquaintance's laptop, and saw that the acquaintance was reading an email message from Jen about her embarrassing experience.
FULL APOLOGY: Later, Colleen ran into Jen and told her what she had seen. Jen admitted to sending the message and said she was sorry.
PARTIAL APOLOGY: Later, Colleen ran into Jen and told her what she had seen. Jen did not admit to sending the message or explain why she had done so, but told Colleen that she was sorry Colleen was feeling bad about what happened.
NO APOLOGY: Later, Colleen ran into Jen and told her what she had seen. Jen did not apologize for what she had done.
A2. Own perspective version of sample vignette
You had an embarrassing experience with a guy from one of your classes--in brief, you thought that he was interested in you, but you were quite wrong about that. You tell two of your close friends (Jen and Allison) about what happened, and ask them to keep it a secret. Two days later, though, you glance at the screen of an acquaintance's laptop, and see that the acquaintance is reading an email message about your embarrassing experience. You can't be sure, but you think you see something that suggests that the message was from Jen. All day long you wonder about whether Jen revealed your secret.
Darby, B. W., & Schlenker, B. R. (1982). Children's reactions to apologies. Journal of Personality and Social Psychology, 43, 742-753. http://doi.org/bt62v2
De Cremer, D., Pillutla, M. M., & Folmer, C. R. (2011). How important is an apology to you? Forecasting errors in evaluating the value of apologies. Psychological Science: A Journal of the Association for Psychological Science, 22, 45-48. http://doi.org/bzcrdw
Eaton, J., Struthers, C. W., & Santelu, A. G. (2006). The mediating role of perceptual validation in the repentance-forgiveness process. Personality and Social Psychology Bulletin, 32, 1389-1401. http://doi.org/b38vh6
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. http://doi.org/csg8r3
Fehr, R., & Gelfand, M. J. (2010). When apologies work: How matching apology components to victims' self-construals facilitates forgiveness. Organizational Behavior and Human Decision Processes, 113, 37-50. http://doi.org/d2ftmc
Goffman, E. (1971). Relations in public. New York: Harper.
Gold, G. J., & Weiner, B. (2000). Remorse, confession, group identity, and expectancies about repeating a transgression. Basic and Applied Social Psychology, 22, 291-300. http://doi.org/fc8d46
Lazare, A. (2004). On apology. New York: Oxford University Press.
McCullough, M. E., Worthington, E. L., Jr., & Rachal, K. C. (1997). Interpersonal forgiving in close relationships. Journal of Personality and Social Psychology, 73, 321-336. http://doi.org/dch699
McFarland, C., & Miller, D. T. (1990). Judgments of self-other similarity: Just like other people, only more so. Personality and Social Psychology Bulletin, 16, 475-484. http://doi.org/dq4tnx
Mikula, G., Scherer, K. R., & Athenstaedt, U. (1998). The role of injustice in the elicitation of differential emotional reactions. Personality and Social Psychology Bulletin, 24, 769-783. http://doi.org/dz2f9g
O'Leary, T. (2008). Five steps to an effective apology. Retrieved from http://www.pickthebrain.com/blog/5-steps-to-an-effective-apology/
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. http://doi.org/bs5p3b
Robbennolt, J. K. (2003). Apologies and legal settlement: An empirical examination. Michigan Law Review, 102, 460-516. http://doi.org/cxw5r7
Robbennolt, J. K. (2006). Apologies and settlement levers. Journal of Empirical Legal Studies, 3, 333-373. http://doi.org/cwmc3c
Robbennolt, J. K. (2008). Attorneys, apologies, and settlement negotiation. Harvard Negotiation Law Review, 13, 349-397.
Roseman, I. J., & Smith, C. A. (2001). Appraisal theory: Overview, assumptions, varieties, controversies. In K. R. Scherer, A. Schorr, & T. Johnstone (Eds.), Appraisal processes in emotion: Theory, methods, research (pp. 3-19). New York: Oxford University Press.
Scher, S. J., & Darley, J. M. (1997). How effective are the things people say to apologize? Effects of the realization of the apology speech act. Journal of Psycholinguistic Research, 26, 127-140. http://doi.org/d84b7t
Schumann, K., & Ross, M. (2010). Why women apologize more than men: Gender differences in thresholds for perceiving offensive behavior. Psychological Science: A Journal of the Association for Psychological Science, 21, 1649-1655. http://doi.org/fvwvj4
Skarlicki, D. P., Folger, R., & Gee, J. (2004). When social accounts backfire: The exacerbating effects of a polite message or an apology on reactions to an unfair outcome. Journal of Applied Social Psychology, 34, 322-341. http://doi.org/bn7dw8
Trope, Y. (1986). Identification and inferential processes in dispositional attribution. Psychological Review, 93, 239-257. http://doi.org/dhchqt
Trope, Y., Cohen, O., & Alfieri, T. (1991). Behavior identification as a mediator of dispositional inference. Journal of Personality and Social Psychology, 61, 873-883. http://doi.org/cfnfg8
Walsh, E. (2009, November 18). Did Goldman's apology backfire? Podium Pundits. Retrieved from http://www.podiumpundits.com/2009/11/18/did-goldmans-apology-backfire/
Worthington, E. L., Jr. (2003). Forgiving and reconciling: Bridges to wholeness and hope. Downers Grove, IL: Intervarsity.
LEONARD S. NEWMAN AND LINDSAY R. KRAYNAK
Leonard S. Newman and Lindsay R. Kraynak, Department of Psychology, Syracuse University.
Correspondence concerning this article should be addressed to: Leonard S. Newman, Department of Psychology, Syracuse University, 430 Huntington Hall, Syracuse, NY 13244-2340, USA. Email: firstname.lastname@example.org