Physical Abuse of Children

An Attributional or Social-Cognitive Approach to Causality

Physical abuse is one of the leading causes of death for children worldwide.  UNICEF (2003) has estimated that, in Organization for Economic Cooperation and Development member states, child abuse and neglect lead to 3500 deaths per year.  Furthermore, the total number of cases of abuse is estimated to be as much as two-thousand times higher than the number of deaths due to maltreatment.

What is the explanation for, or the cause of, child abuse?  Attempts to understand and mitigate child maltreatment have met with limited success.  A number of factors have been identified as correlates of child maltreatment.  Included in these factors are low socioeconomic status, a cultural background tolerant of violence, a breakdown of the family, social isolation, child morbidity, parental mental illness and substance abuse, and parents who were abused during their own childhood.  However, the events that lead to maltreatment are complex and are not well understood within any single theoretical framework.  Frustration with the low impact of etiological research in this area may explain the shift of focus for many researchers, a shift away from causes of abuse toward interventions with abuse survivors.
While it is unlikely that an all-encompassing theory is possible, an attributional approach to advancing the understanding of this multi-faceted phenomenon has gained some currency in psychological literature.  Attribution theory predicts that some form of aggressive behavior, such as child abuse, will be focused on the person or object perceived to be the intentional cause of a negative event.  In contrast, persons or objects perceived as causally-linked with a negative event but in an unintentional fashion are less likely to be the recipients of focused aggression; in fact, a sympathetic response is posited as more probable under these conditions.

The following is a discussion of the application of this model to the domain of child maltreatment; in particular, physical abuse.  In the specific model to be discussed, physical abuse is conceptualized as an instance of aggression, and attributional processes are imbedded within the context of a social-cognitive approach to aggression.  The utility and limitations of this theoretical framework, and its implications as a model for preventive intervention, will be discussed. 
It should be understood at the outset that donning any one particular theoretical perspective, with its defined and finite set of concepts, constructs and relationships, imposes a necessarily limiting condition on understanding causality as it relates to the breadth and scope of the child abuse phenomenon.  Perforce, a multitude of other contributing variables is disregarded.

About the Model

With a focus on physical child abuse as an incidence of aggression, it first is necessary to understand the social-cognitive approach to understanding aggression.  The somewhat mechanistic frustration-aggression hypothesis (Dollard et al., 1939), which dominated earlier research on aggression, alerted the researcher and practitioner to look for a prior frustrating event.  This approach had, and continues to have, advocates and applicability in the field of child abuse research.

By contrast, an implication of the social-cognitive approach is that, given the presence of aggressive behavior, one should look for the aggressor’s judgment that the victim is personally responsible for a prior negative event and for the aggressor’s feelings of anger toward the victim.  This model assumes that the perceptions of negative events, inferences regarding potential motives for those events, and other information-processing activities are central to understanding the instrumentation and etiology of hostile behaviors, the how and the why of aggression (Crick & Dodge, 1994).  Within this perspective, attributional analyses focus on the process of attributing or assigning causality for the occurrence of effect events that are experienced as positive or negative, in varying degrees, by the perceiver.  Attributional models differ from other social-cognitive forms of analysis in the inclusion of affect as a central construct (Weiner, 1986, 1995).  Adding affect to the mix results in a broadening of the quality, as well as quantity, of potential predictors of aggression.  As Weiner (1995) pointed out, affect takes us beyond the realm of cold and objective cognitive factors.

In attribution theory, the perceived intentionality/controllability of an effect event experienced as hostile constitutes the deciding factor for the prediction of whether an aggressive response will occur.  If a person attributes a negative event to the voluntary actions of another, then some form of aggression directed at the perceived cause of the event can be predicted.  On the other hand, if the perceived cause of the negative event is believed to have acted involuntarily, then a less aggressive response would be predicted.  Whether the causal act is perceived as within or beyond the actor’s control is pivotal here, and the assignment of responsibility for the act determines the quality of the response. 

Injurious actions, for example, elicit inferences of responsibility if the actor is perceived to possess knowledge of the nature of the act and the intent to inflict injury.  Conversely, responsibility is less likely to be assigned to an actor whose behavior is deemed to be outside his control or the negative affect associated with the action is judged to be unintended (Graham, Weiner, & Zucker, 1997).  A notorious illustration of this mediational role for volition/controllability in the assignment of responsibility may be seen in the Nazi war crimes adjudication proceedings.  A less extreme example might be the greater assignment of responsibility for financial success to a person perceived as having worked hard than to a lottery winner.

These judgments of volition and intent are central to the prediction of aggression from an attributional viewpoint.  An often-replicated finding in the literature on childhood aggression is that children and adolescents who tend to exhibit aggressive behavior are more likely to assign hostile intentions to others than are their less aggressive peers (Crick & Dodge, 1994).  That perception can constitute a motive for further aggression in the guise of retribution and justice.
In addition, Averill (1983) and Weiner (1995) provided evidence demonstrating that the perception of personal responsibility for an injurious act can evoke anger and the related, affectively-negative experiences. This finding is in line with the idea from appraisal theory that thoughts have the capacity to elicit emotions (Ellsworth & Smith, 1988).  In turn, anger has been demonstrated to arouse aggressive, retaliatory behavior (Berkowitz, 1993).  The physiological experience of anger can function as a stimulus to hostile action.

Judgments regarding responsibility for an effect event also can result in more positive resultant behavior.  Contextual cues provide a rich source of information for the assessment of personal responsibility.  For example, if one’s comment is ignored by another, the negative affect, anger and aggression that might be generated given the assumption of an intentional snub could be mitigated by the explanatory cue of a noisy room. If an individual is not held personally responsible for causing a negative event, then the door may be opened for a sympathetic response to that person (Schmidt & Weiner, 1988; Weiner, 1995).

In fact, studies of altruism (i.e., helping behavior) have provided very strong support for the predictive value of attributional approaches.  This literature testifies to the role of inferred responsibility in mediating behavioral responses to the perceived cause of an effect event. It has been demonstrated, time-after-time, that people tend to respond with sympathy and altruistic behavior given that the person in need of help is not judged to be responsible for his/her plight. Conversely, if the cause of a person’s distress is attributed to actions within the person’s voluntary control, then people tend to respond with anger and to withhold help.  For example, the tendency to come to the aid of a student, on crutches and wearing a cast, who drops an armful of textbooks, should be more pronounced than helping behavior manifested toward the same student’s dropping a pile of magazines that extol extreme sports.

Just as attributional processes do not mediate all aggression, altruism may be found in the absence of responsibility-mediated attributions.  However, attribution theory has a significant distinction in its ability to be applied to, and have predictive validity within, the domains of both pro-social and antisocial interpersonal behavior.  For many researchers, this evidence of the rich and robust quality of the attributional framework positions it among the general theories of human motivation.

Applying Attribution Theory to Child Abuse

Two sequences characterizing the etiology of aggressive versus nonaggressive responses to negative effect events can be derived from the theoretical relationships and the empirical evidence cited thus far:

• Attribution of causality for a negative effect event to a particular person (target)  inference of personal responsibility for the negative event  increased anger and decreased sympathy  aggressive behavior directed toward the target.

• Attribution of causality for a negative effect event to a particular person (target)  no personal responsibility for the effect event is inferred  decreased anger and increased sympathy  no aggressive behavior directed toward the target.

These attribution-assessment-emotion-behavior sequences can be applied specifically to the realm of physical child abuse.  Consider the following scenario.  Little Janey tracks mud into the house after playing outside.  The mother knows that Janey is the cause for muddy footprints on her clean floor (i.e., attribution of causality for a negative effect event to a target).  The mother believes that Janey did this intentionally, to cause her more work (i.e., locates personal responsibility for the event in the target).  The mother becomes angry and strikes Janey.  Alternatively, the mother may realize that Janey did not purposively muddy the floor; that she was attempting to honor her newly-taught potty training by rushing to the bathroom.  The muddy floor remains a negative event and Janey’s behavior remains the cause; however, the intention to perform a negative action is not assigned to Janey.   Therefore, the mother directs less anger and more sympathy toward her daughter, and aggressive behavior toward Janey is not the chosen response.

The determinants of child maltreatment include both attributional (cognitive) and affective (emotional) components.  This interpretation of the causes for physical child abuse has received some support in the research literature.  A key antecedent of abuse was, at one time, believed to be unrealistic expectations on the part of the parents regarding the developmental pace of the child (Spinetta & Rigler, 1972).  These false expectations can be interpreted as inferences of controllability that mediated aggressive responses (e.g., Bradley & Peters, 1991).  Other findings related to this point have indicated that abusive parents tend to perceive intentionality or control by the child in the performance of negative behaviors (Bugenthal, 1987; Bugenthal et al., 1989; MacKinnon-Lewis et al., 1992).  These data are consistent with the attributional analyses of child abuse reported by Bauer and Twentyman (1985), Graham and colleagues (2001), and Larrance and Twentyman (1983).  Milner and Foody (1994) reported the resistance of parents at-risk for child abuse to changing their attributions of intentionality on the part of the child, even in the face of mitigating information (e.g., contextual cues). 
In summary, there are empirical findings in support of an attributional approach to understanding physical child abuse.  However, the number of studies is relatively small.

Implications for Intervention

The attributional approach and research findings reviewed have implications for preventive interventions with at-risk caregivers.  One obvious starting point is attributional change, training caregivers to see their children as less in control of, and less responsible for, their negative behaviors.  Attributional therapy has been used to produce changes in behavior by altering causal thinking in educational and clinical settings for more than twenty years (Forsterling, 1985).  Abusive caregivers can be instructed regarding the meaning responsibility, how accurately to infer intentionality, and how circumstances can change inferences regarding blame.

Conclusion

A shared belief among researcher and theorists is that multiple sufficient causes exist and apply to particular manifestations of aggression, including child abuse.  For example, Belsky’s (1993) review of the literature dealing with child maltreatment concluded that: “All too sadly, there are many pathways to child abuse” (p. 413).  Were your parents abusive?  Is it too warm in your house?  Are you prejudiced?  Do you feel frustrated?  Are you mentally ill?  A “yes” response to any of these questions, as well as a score of others, indicates that you are, to some degree, at risk for abusing a child. 
Any single account can encompass only a small portion of the dynamics of aggression, in general, or of child abuse, in particular.  While the telling of a coherent tale is the virtue of adherence to a particular theoretical framework or path, the path itself, by rights of its own particular attributes, imposes limits of understanding on the story-teller and the audience.

Attribution theory promotes a compelling view of the cognitive and affective factors that can lead to physical child abuse or to a sympathetic response.  Perhaps it is most compelling in its offering of interventions to prevent mistreatment of children, a virtue only briefly touched upon in this paper. 

Overall, more research is needed, as well as the recognition that the part of the etiological story based on attribution theory is a small part, indeed.