The Future of Children, Princeton - Brookings: Providing research and analysis to promote effective policies and programs for children.

In 1954, Cyril Burt introduced the fourth edition of his The Young Delinquent with the following authoritative statement: "Thirty years ago, when I first began my work, the most urgent need appeared to be a systematic and scientific investigation into the chief causes of delinquency. Now, as we have seen, a general measure of agreement has at length been reached [as to both] the nature and the relative importance of the various causal factors."¹⁸

Burt thought the cause was the psychological traits or inherent defects of individuals, but other researchers have believed, with equal conviction, that the root causes of violence were environmental. Jack Tizard, for example, suggested: " . . . we need to develop a theory about the environmental determination of characteristic ways of behaving in different circumstances . . . and research on the characteristics that differentiate one natural environment from another, and to explore ways of changing some of them and of assessing their effects."¹⁹

Still other researchers have posited that genetic factors, neurochemical mechanisms, temperament, family environment, early experiences, institutional settings, social conditions, or labeling were important determinants of violent behavior.²⁰ Currently, the most hotly debated theory regards the potential role of genetic and neurochemical factors such as the gonadal hormone, testosterone, and the neurotransmitter, serotonin. While evidence exists for a correlation between these substances and aggression, the findings are hardly conclusive.²¹ It may be that preexisting behavior or other environmental conditions determines the level of such substances. For some biological systems, biological factors may be the consequences, not causes, of behavior.²²

Any proposed theory must fit the data. That is, an adequate theory must explain why rates of violence differ by age, gender, and size of community, and whether it is environmental conditions that cause individuals to behave violently (social causation), or whether persons prone to behave violently create environments characterized by high rates of violence (social selection).

Theories must fit one other key finding that has emerged from the literature on root causes of violence: some children show signs of disruptive and antisocial behavior from as early as 2 to 3 years of age.²³ These children are likely to continue to show signs of disruptive behavior throughout childhood, and to engage in a higher frequency and more severe delinquency during adolescence than others. Theories should therefore explain the backgrounds and causal pathways that distinguish this important group of delinquent youths because it is this group that is likely to be disproportionately involved in offending.²⁴

Developing a theory that fits the data is a tall order, and no theory presently comes close to approximating these requirements. As suggested by Tizard, the difficult task that remains is development of an integrative theory—one that deals simultaneously with individual differences and contextual influences.

Typically, theories are tested by examining rates of crime in large groups and seeing if models can be built to predict which children will display violent behavior later in life, or which adolescents or adults may, as young children, have had particular experiences that predisposed them to later violence. The key experiences or characteristics of individuals that are used in these models are called risk factors.

Investigators typically select risk factors to include in a model based on a favorite theory. For example, disruption in mother-infant bonding might be the most important risk factor included in a model by a psychoanalyst, but exposure to violence in the home and inconsistent disciplinary practices might be the risk factors included in a model by a social learning theorist. Eventually, however, most researchers try to include as many variables as feasible. What may have begun as a theoretically guided investigation therefore can sometimes turn into an exploratory exercise. This is probably an inevitable response to the nature of violent behavior. No one theory or limited set of risk factors can go very far in producing a powerful explanation, given the multiply determined nature of violent behavior.

Understanding the different kinds of risk factors is essential to interpreting how violent behavior is caused. Variables such as sex, race, ethnicity, and social class, for example, are sometimes called risk factors for violent behavior. Such variables are most usefully employed to increase precision in identifying and locating vulnerable groups. They may be misused if they are assumed to be part of a mechanism that causes delinquency. To be clear, it is perhaps best to refer to these noncausal types of risk factors as population indicators. In contrast, other risk factors such as temperament, family relationships, and school performance do suggest the operation of potential causal mechanisms.

It is also important to consider the multiple levels on which risk factors typically operate. Risk factors can exert influence on individual, family and peer, institutional, community, or societal levels. Psychologists and psychiatrists often build theories beginning with individual factors and gradually incorporating environmental influences. Sociologists, on the other hand, begin with societal and institutional influences and gradually incorporate individual-level factors. Psychologists and sociologists have rarely engaged in interdisciplinary efforts in this field. Yet, it is during such encounters that the search for understanding may be most powerfully advanced.

Risk factors for juvenile violence may be organized into the following five domains: (1) social disorganization of communities, (2) poor school climate, (3) exposure to deviant peers (through gang membership and delinquent behavior, access to weapons, and substance use), (4) adverse family relationships, and (5) individual physical or psychological predispositions. Box 1 outlines some of the variables that represent these domains and that have been demonstrated through a variety of studies to be the most important characteristics associated with juvenile violence.

Some risk factors are more important than others, although their relative importance may shift with age and maturation. Relevant studies to address questions about the magnitude of the relationship between risk factors and violent behavior are limited. Nevertheless, based on the strong assumption that multiple developmental pathways lead to violent behavior—one for those who display chronic violent behavior, beginning at an early age, another for those who display violent behavior only later in life, and others in which there exists no substantial evidence of previous problem behavior—it can be hypothesized that the risk factors and causal mechanisms will be different for these groups. Using the list provided in Box 1, we might hypothesize that family adversity and individual psychological and health-related problems are most closely associated with violence in the early-onset, high-persistence pathway, while community and peer influences are most important for the more late-onset and transient pathway.²⁶

It is common in developmental psychology today to cite protective factors to explain why only some children living in adverse conditions develop problematic outcomes. Perhaps the children who escape adverse consequences are exposed to or possess a set of protective factors. Research about protective factors is still at an early stage, but is summarized in the recent Office of Technology Assessment (OTA) report on adolescent health as follows: "Overall, a picture emerges of the resilient child as having an easy temperament and a higher IQ, being more autonomous, having a good relationship with at least one adult, and being more successful and involved in school."²⁷

Results of studies of risk and protective factors should be applicable to the design of interventions, which is discussed in the following section. But, as pointed out in the OTA report, crafting an intervention is difficult because, " ...few, if any, risk factors for delinquency act independently. Many of the risk factors for adolescent delinquency are interrelated in ways that are still not well understood. It is clear that no one factor by itself is correlated with or predicts delinquency very well, but rather for most adolescents, delinquency is the result of the interaction of multiple risk factors..., each of which incrementally increases the risk of delinquent behavior. The importance of each factor also depends on the age of the individual."²⁸ [Italics in the original.]