The effects of empathy and outcome on attribution: An examination of the divergent-perspectives hypothesis

Based on the theoretical ideas of Jones and Nisbett (Jones et al. Attribution: Perceiving the cause of behavior. New York: General Learning Press, 1971), and the recent findings of Regan and Totten (Journal of Personality and Social Psychology, 32, 1975, 850–856), the present study assumed that from an attributional standpoint empathic observers and actors are functionally equivalent. On this basis it was predicted that empathic, relative to nonempathic, observers would make outcome attributions which have been typically found for actors themselves: They would attribute an actor's success to dispositional causes, but an actor's failure to situational causes. After instructions to empathize with the target, or to observe him, subjects watched a videotape of a target male attempting to make a good first impression on a female. Subjects later learned that the target had either succeeded or failed at making a good first impression, and were asked to make causal attributions for his outcome. As predicted, instructions to empathize led to dispositional attributions for success and situational attributions for failure, while standard observation instructions resulted in dispositional causal attributions regardless of outcome. The results were interpreted as supporting the contention that differential information processing may sufficiently account for the effects of outcome on causal attributions.     

Esteem and control bases of blame: "Adaptive" strategies for victims versus observers

An experiment was conducted to determine the esteem and control correlates of behavioral and characterological blame for victims and observers. On the basis of a proposed motivation to minimize perceptions of vulnerability, it was predicted that behavioral self-blame would be “adaptive” for victims, whereas both behavioral and characterological blame of the victim would be “adaptive” for observers. As hypothesized, behavioral self-blame by victims was associated with high self-esteem and perceptions of future avoid ability of the victimization, whereas both behavioral and characterological blame by observers were associated with high self-esteem. Behavioral blame was more likely to be engaged in by victims than by observers, and characterological blame was more likely to be engaged in by low esteem victims and high esteem observers. The study employed a role-playing/observer methodology. The adequacy of this methodology and the generalizability of results to actual victim/observer populations are discussed.     

Empathy and attribution: turning observers into actors.

A laboratory experiment was conducted to test Jones and Nisbett's information-processing explanation of the often-observed tendency for individuals (actors) to provide relatively more situational and less dispositional causal attributions for their behavior than those provided by observers of the same behavior. According to this explanation, aspects of the situation are phenomenologically more salient for actors, whereas characteristics of the actor and his behavior are more salient for observers. To test this explanation, the phenomenological perspective of observers are altered without making available any additional information. Subjects watched a videotape of a get-acquainted conversation after instructions either to observe a target conversant or to empathize with her. As predicted, taking the perspective of the target through empathy resulted in attributions that were relatively more situational and less dispositional than attributions provided by standard observers. The results support Jones and Nisbett's information-processing explanation of actor-observer attributional differences, and shed additional light on the process of empathy.     

Turning observers into actors: Differential causal attributions as a function of “empathy”

After being instructed either to “empathize with the actor” or to “picture the events clearly,” two groups of observers read a story describing an actor's behavior, and then gave free-response explanations of that behavior, and rated the importance of personal and situational causal factors. The hypothesis that causal attributions of empathizing observers would be less personal and more situational than those of nonempathizing observers received strong support, both from subjects' free responses and from their scale ratings. These findings provide evidence for an information-processing explanation of actor/observer attributional differences. Some practical applications of increasing the situationality of observers' causal attributions are discussed. The results also suggest a novel operational definition of “empathy”; and are interpreted as evidence for the effectiveness of “interpersonal simulations”.     

Women's Reactions to Rape Victims: Motivational Processes Associated With Blame and Social Support1

A study was conducted with 128 female college students to test the hypothesis that when observers feel vulnerable to rape, they are more likely to blame a rape victim³ and are less willing to offer social support. Similarity and empathy were expected to moderate the effects of perceived vulnerability on blame and predict greater social support. Assumptions about the world were predicted to be associated with greater blame. A multivariate model was tested with structural equation modeling techniques. Perceived vulnerability did not directly or indirectly predict blame. However, similarity directly predicted less blame and indirectly predicted greater social support through associations with blame, perceived vulnerability, and empathy. World assumptions directly predicted greater blame and indirectly predicted less social support through blame. These findings suggest that blame and social support are interrelated processes which are associated with social observers' perceptions of the victim and their basic assumptions about the world.     

Perceptions of adolescent bullying: Attributions of blame and responsibility in cases of cyber‐bullying

Cyber‐bullying (where victims are targeted via online social networking or other electronic means) has gained increased attention in research and the broadcast media, but previous research has not investigated attribution of blame in such cyber‐bullying events. This experiment hypothesized that participants would assign higher ratings of blame to bullying perpetrators when the bullying situations were depicted as having highly foreseeable outcomes (vs. unforeseeable outcomes), and as occurring in school (vs. online). In addition, a significant interaction was predicted between outcome foreseeability and bullying situation, with highly foreseeable in‐school events being rated as the most predictable and attributable to the bully's actions. One‐hundred sixty‐three participants completed surveys containing demographic items, items regarding their past experiences of victimization, and one of four randomly‐assigned vignettes detailing a bullying situation (which participants rated). While hypotheses regarding outcome foreseeability were supported, no cyber‐bullying vs. in‐school main effects (or corresponding interaction effects) were detected. Implications for future research and practice, as well as study limitations, are discussed.     

Perceived Responsibility and Blame for Economic Success and Failure: Social Class and Employment Status Comparisons

The present study tested the proposition that social class and employment status would differentially affect attributions of responsibility and blame concerning positive and negative economic outcomes. The hypothesis here is that subjects will not blame a person who fails to get a job after taking a retraining program. The hypothesis is based upon the assumption that individuals on the lower rung of the economic ladder are aware of the external circumstances operating against economic security. As predicted, the person experiencing failure was neither held responsible nor blamed for his outcome. Also social class and employment status affected perceptions of the reasons for the successful outcome.     

Subjective probabilities and moral judgments

The two experiments reported here examined the relationship between subjective probability estimates and moral judgments (credit and blame assignment, trait attributions, and behavior evaluations). Subjects read about situations that varied in outcome valence (moral or immoral); in addition, the nature of situational demands (Experiment 1) or behavior frequency (Experiment 2) was varied. In the first experiment, subjective probabilities were related to judgments of immoral behaviors (but not moral behaviors), whereas the situational demands only had an impact on judgments of moral behaviors. Experiment 2 included a wider range of behavioral situations, and the probability estimates and moral judgments were assessed independently. In contrast to the first experiment, subjective probabilities were related to trait and behavior ratings of both moral and immoral acts. Consistent with the first experiment, however, subjective probabilities predicted blame but not credit. Across both studies, the prior expectancies were more strongly related to evaluations of immoral acts than moral acts. Implications for understanding the determinants of judgments of moral and immoral acts are discussed.     

Observer Perceptions of Would-Be Provider and Would-Be Recipient Reactions to Rejected Social Support1

Role-playing observers predicted the reactions of hypothetical would-be recipients whose requests for social support were unmet as well as those of hypothetical would-be providers for their rejection of such requests. The extent to which the would-be recipient was depicted at fault for the problem was predicted to affect perceived reactions since support may be less likely when potential recipients are blameworthy than when they are innocent of responsibility for their predicament. Observers perceived would-be recipients of support as believing that would-be providers blamed them. Observers saw would-be providers as blaming the would-be recipients rather than themselves, especially if the victim was blameworthy. When the victim was not blameworthy, would-be providers were more likely than would-be recipients to be judged to use a mixture of rationalizations and excuses rather than blame.     

Culpable control and counterfactual reasoning in the psychology of blame

Many counterfactual reasoning studies assess how people ascribe blame for harmful actions. By itself, the knowledge that a harmful outcome could easily have been avoided does not predict blame. In three studies, the authors showed that an outcome's mutability influences blame and related judgments when it is coupled with a basis for negative evaluations. Study 1 showed that mutability influenced blame and compensation judgments when a physician was negligent but not when the physician took reasonable precautions to prevent harm. Study 2 showed that this finding was attenuated when the victim contributed to his own demise. In Study 3, whether an actor just missed arriving on time to see his dying mother or had no chance to see her influenced his blameworthiness when his reason for being late provided a basis for negative evaluations but made no difference when there was a positive reason for the delay. These findings clarify the conditions under which an outcome's mutability is likely to influence blame and related attributions.     

Moral judgment and the development of causal schemes

The present study investigated (i) the relationship between blame and perceived causality; (ii) the effect of the nature of causes on causal inference. Seventy-two persons from three age groups (5, 9 years and adults) responded to behavioural events which varied in outcome intensity, the nature of the cause (internal/external) and its presence (present/absent). The latter two factors had a marked effect on attributed blame and inferred causes as an age × nature × presence of cause interaction was found in both cases. However, inferred causes were not systematically related to attributed blame. Outcome severity led to more extreme blame ratings in all groups but only affected the causal scheme used by adults. The results are discussed in terms of over-attribution to persons and a more precise criterion for the use of the multiple sufficient cause scheme is evaluated.     

Causal attributions for behaviors consistent or inconsistent with an actor's personality traits: Differences between those offered by actors and observers

The present research was designed to investigate differences in the attributions offered from the actor's perspective and the observer's perspective. It was predicted that causal attributions for behaviors inconsistent with an actor's personality traits would be more situational when offered from the actor's perspective than when offered from the observer's perspective. In contrast, it was predicted that causal attributions for behaviors consistent with an actor's personality traits would be more dispositional when offered from the actor's perspective than when offered from the observer's perspective. Consistent with these hypotheses, extraverts explained introverted behaviors and introverts explained extraverted behaviors more situationally from the actor's perspective than from the observer's perspective. Furthermore, extraverts explained extraverted behaviors and introverts explained introverted behaviors more dispositionally from the actor's perspective than from the observer's perspective. These differences in the attributions offered by actors and observers were attenuated but not eliminated when attributors had access to useful situational information with which to apply the discounting principle.     

Causal attribution of success and failure as a function or authoritarianism and sex.

We predicted that authoritarian actors would engage in defensive attribution, and authoritarian observers would derogate the other, to a greater extent than egalitarian perceivers. 48 male and 48 female college students were run in pairs of same sex and authoritarianism. A set of easy anagrams was given to subjects in success conditions and difficult anagrams to those in failure conditions. Each subject rated own outcome and other outcome in terms of internal factors of ability and effort, and external factors of task and luck. We found that authoritarian actors were more internal than egalitarian actors only in the condition own success-other's failure. Authoritarian observers, as compared to egalitarians, were more external for other's success and more internal for other's failure only when own outcome was successful. It seems that authoritarian perceivers exaggerate their abilities and derogate the other only when they are clearly in a superior position vis-a-vis the other. There were no sex differences as a function of outcome and authoritarianism.     

That Could Be Me (or Not): Senseless Violence and the Role of Deservingness,Victim Ethnicity,Person Identification,and Position Identification1

Building on just‐world theory, the current study examined variables contributing to the labeling of violent incidents as senseless. In a 2 × 2 (Blame Opportunities x Victim Ethnicity) design, Dutch participants (N= 78) were provided with a written hypothetical situation depicting a violent incident. Consistent with predictions, the violence was evaluated to be less deserved and more senseless (and the desired penalty for the offender was stronger) when participants had no opportunity to blame the victim than when they did have an opportunity to blame the victim. Likewise, an act of violence committed against a victim belonging to an ethnic minority (allochthonous victim) was perceived to be more deserved and less senseless (and the desired penalty for the offender was smaller) than a similar violent act directed against a native (autochthonous) victim. Findings designate that the just‐world theory offers a promising approach to investigate factors determining the labeling of violent incidents as senseless by outside, uninvolved observers.     

Conditioning of counter-empathetic emotional responses

The present research tested the hypothesis that the quality of an observer's vicarious emotional response, as measured by autonomic, expressive, and self-report indexes, is a function of the observer's conditioning history with particular facial expressive displays of emotion. It was predicted that conditions of congruence (Symmetry) between the affective expression of a model and the outcome (shock or reward) presented to an observer would enhance initial empathetic responses, but that conditions of incongruence (Asymmetry) between the model's displays and observer's outcomes would lead to counter-empathetic responses. These changes in the quality of observers' vicarious emotional responses should generalize to a test phase when no rewards or punishments are presented to observers. The results for all measures were consistent and indicate that asymmetric conditioning modified the initial empathetic responses of observers to either counter-empathetic responses or indifference. On the other hand, symmetric conditioning enhanced observers' initial empathetic responses. These effects were evident in the test phase when no reinforcements were administered to the subject. The results are consistent with the theoretical assumption that facial expressions of emotion can acquire meaning and hedonic valence because of their predictive significance and thus can function as conditioned stimuli capable of evoking empathetic and counter-empathetic emotional responses.     

Figural emphasis and person perception

Based on Jones and Nisbett's (1972) proposition that actor-observer differences in causal attributions derive from differences in attentional focus, it was hypothesized that observers' focus of attention would influence their causal attributions for an actor's behavior. More specifically, it was predicted that the behavior of an actor who was the focus of attention by virtue of some salient physical attribute would be attributed by observers more to dispositional causes and less to situational causes than would the behavior of a less physically salient actor. The manipulations of physical salience were based upon Gestalt laws of figural emphasis in object perception. They included brightness (Study I), motion (Study II), pattern complexity (Study III), and contextual novelty (Studies IV and V). The results revealed that the salinece of the actors' environments (i.e., the other people present) rather than the salience of the actor him/herself had the most consistent influence on causal attributions. When environmental salience was high, behavior was attributed relatively more situationally than when it was low. Prior research findings are considered in light of the proposition that causal attributions for an actor's behavior vary only with the salience of his/her environment, and additional implications of this phenomenon are suggested. Some ambiguities in the application of Gestalt principles to the perception of people are discussed.     

Dependency and fault as determinants of helping

In a natural setting, helping was investigated as a function of a female requestor's dependency on the subject for help and whether or not she was at fault for her plight. Based on previous research and the hypothesized operation of two norms, it was predicted that the negligent requestor would be more likely to receive help than the victim of circumstance when her dependency was high, but would be less likely to receive help when her dependency was low. Randomly selected telephone subscribers received a “wrong number” telephone call from a stranded women motorist. The woman asked the subject to make a phone call for her; the dependent variable was whether the subject helped by making the call. Results were consistent with predictions. When dependency is high, the victim's negligence appears to operate as an indication of greater need, whereas when dependency is low, it seems to operate as a sign that the victim is less deserving of help. Results were discussed in relation to different social norms that may be activated as a function of the dependency of the requestor.     

Blame contagion: The automatic transmission of self-serving attributions

When people blame others for their mistakes, they learn less and perform worse. This problem is magnified when blame becomes embedded in the shared culture of groups and organizations. Yet, little is known about whether—and, if so, how—the propensity to blame spreads from one person to another. Four experiments addressed this issue, demonstrating that blame is socially contagious: observing an individual make a blame attribution increased the likelihood that people would make subsequent blame attributions for their own, unrelated, failures (Experiments 1, 2, and 4). Results also indicated that this “blame contagion” is due to the transmission of goals. Blame exposure led to the inference and adoption of a self-image protection goal (Experiment 3), and blame contagion was eliminated when observers had the opportunity to alleviate this self-image protection goal via self-affirmation (Experiment 4). Implications for research on causal attributions, social contagion, and cultural transmission are discussed.     

Aggression and power restoration: The effects of identifiability and timing on aggressive behavior

The study was designed to demonstrate that one function of aggression is the restoration of power. It was predicted that, after an individual had his power reduced, greater aggression would be emitted when the individual could be identified as the attacker by his victim than when he could not. It was felt that only when the aggressor was identifiable could he completely restore his power vis-à-vis the victim. A second aim of the study was to investigate the effects of timing of aggression on the intensity of aggression. It was predicted that in a learning situation, if aggression were utilized solely as a teaching device, greater aggression would be emitted by the “teacher” when the “learner” made mistakes early as opposed to late in the task. However, if the individual were motivated to restore his power by aggressing, greater aggression would occur when the mistakes were made late since having to wait should frustrate the teacher's desire to restore power through aggression. The design of the study was a 2 × 2 × 2 factorial in which subjects were either insulted or not, given the opportunity to aggress either anonymously or when clearly identified, and able to aggress either early or late in a learning task. The results supported the power-restoration hypothesis: Subjects who were insulted aggressed more when identifiable than when anonymous. Also, there was greater aggression in the early as opposed to late conditions under all circumstances except in the insultidentifiable condition.     

Interpersonal risk and the evaluation of transgressions in close relationships

An experiment examined individuals’willingness to excuse a romantic partner of blame for a transgression when perceptions that a relationship is risky are salient. Participants evaluated an actual transgression on measures tapping three levels of appraisal: (a) initial impressions of the act (i.e., severity of the transgression), (b) considerations of the context in which it occurred (i.e., judgments about excuses and extenuating context), and (c) judgments about its broader implications for the relationship (attributions of globality). Evaluator perspective was also varied. Half the participants (actors) evaluated their own partner's wrongdoing; half (observers) evaluated another participant's partner's wrongdoing. Compared to controls, risk participants rated the transgression as more severe and were more cautious and risk-averse in assessing the merits of potentially excusing information. Evaluator perspective did not influence these judgments, a finding consistent with a cognitive interpretation of the results. In contrast, the effects of risk on judgments of globality were more pronounced among observers than among actors, suggesting that motivational pressures come into play when the evaluative stakes are higher.