Understanding and Preventing Violence, Volume 2: Biobehavioral Influences (1994)
Chapter: Measuring the Dependent Variable
investigations appear incompatible and/or inconsistent. That is to say, we study attack and defense mechanisms in normal cats (and other animals) and violent behavior in abnormal (presumably) human subjects. Are there really any links between the two bodies of research? We must assume that the mechanisms and structures involved in the final common neural pathway for the expression of attack and defense are basically similar in all higher mammals. A further assumption is that the mechanisms for activating or suppressing the aggression/violence pathway(s) are vastly more complex in humans because of the increased size of the forebrain. Nevertheless, it seems reasonable to hypothesize that humans who are engaging in maladaptive, violent behavior have some brain abnormality that somehow temporarily (and, possibly, repeatedly) short-circuits the aggression/violence pathways leading to truly violent behavior. We consider the evidence bearing on this hypothesis in the review and discussion below.
To return to the issue of defining adaptive versus maladaptive aggressive behavior in humans, there is on the one hand likely to be general agreement that serial killers or sexual offenders who maim their victims are displaying disordered, maladaptive violent behavior. On the other hand, the label of "adaptive" would be applied to the behavior of soldiers who kill enemy personnel in the course of battle and to the actions of skilled boxers who inflict physical punishment on (and occasionally kill) an adversary. Unfortunately, most of the examples of human violence and aggression considered in this section appear to fall somewhere between these two extremes. Violent or aggressive behavior may be quite adaptive in young male persons raised and living in impoverished environments; it may also be one of the few means of expression available to persons with impaired cognitive or language capacities. Surveys of violence and aggression have implicated age, gender, socioenvironmental, and/or cognitive variables; it may be virtually impossible to identify the portion of the variance that is attributable to adaptive, as opposed to nonadaptive, factors. Nevertheless, to the extent possible we will try to bear this distinction in mind in the course of reviewing and discussing the areas of human violence.
Measuring the Dependent Variable
The end points in measuring aggression/violence in animal models (see above) are simple, highly reliable, and almost stereotypic motor behaviors. By contrast, the measurement of human
