انتقال بین نسلی رفتار ضداجتماعی: کودکان چگونه به بزرگسالان ضد اجتماعی تبدیل می شوند؟

Abstract Exposure to many potential environmental risk factors for child antisocial behavior is associated with one of the strongest predictors of antisocial behavior, a family history of antisociality. Because most studies of putative environmental factors do not take into account genetic propensities for antisocial behavior shared between parent and child, the possibility of genetic contributions to these “environmental” markers is typically not evaluated. In this paper, we review research on the environmental correlates of antisociality, their association with parental antisociality, and highlight findings from studies that have controlled for either genetic propensities or parental antisociality.

1. Introduction Antisocial behaviors in adulthood are problematic at both the individual and societal levels. At the individual level, antisocial adults tend to have unstable relationships that are marked by conflict. Antisocial adults are also less likely to marry, and those who do marry experience higher rates of domestic violence, divorce, and fatherless children rearing (Cloninger, Bayon, & Przybeck, 1997). At the societal level, antisocial adults are more likely to have unstable work histories and to have multiple arrests and felony convictions (Cloninger et al., 1997). They are more likely to be involved in fights and to use weapons in these fights (Farrington, 1995). Antisociality is both highly problematic and far from rare. While estimates of the prevalence vary, it is safe to say that antisocial behavior is common. Studies of antisociality have examined several different phenotypic expressions of problematic behaviors (e.g., conduct disorder (CD), antisocial personality disorder (ASPD), oppositional defiant disorder (ODD), delinquency, criminality, aggression, and violence), which we will refer to collectively as antisocial behavior. While each of these phenotypes of antisociality describes different specific behaviors, there is reason to believe they are nonetheless etiologically related. Krueger et al. (2002) reported that high rates of comorbidity among conduct disorder, adolescent antisocial behavior, and a disinhibited personality style are due to these behaviors loading on a shared latent externalizing dimension (see also Kendler, Prescott, Myers, & Neale, 2003; Young, Stallings, Corley, Krauter, & Hewitt, 2000). Additionally, while many behaviors can lead to criminal conviction, there is evidence that delinquency itself is a part of a unitary factor (Farrington, 1995). Thus, it appears that various antisocial characteristics often examined separately frequently co-occur and may be manifestations of a unified domain of human individual differences. Therefore, in this review we do not attach special significance to any one of these many manifestations of antisociality and will not limit ourselves to one particular operationalization of antisocial behavior. Rather, this paper will focus on a range of destructive forms of antisocial psychopathology including delinquency, criminality, violence, CD, ASPD, and aggression. Despite the individual and societal toll exacted by antisocial behaviors, factors contributing to the development of antisocial disorders are not well understood. While a number of familial factors, such as divorce and maladaptive parenting, have been linked to child antisociality, these factors are also associated with parental antisocial behavior. As antisocial behaviors tend to run in families and to be partially heritable, it is unclear to what extent antisociality in the parents may be leading both to child antisocial behavior and to exposure to the identified risk factors. For example, it is possible that divorce may only be associated with heightened rates of child antisocial behavior because antisocial adults are more likely to divorce, and thus may be overrepresented in divorced samples. The children of these antisocial parents are at an increased genetic risk for antisocial behavior. Higher rates of child antisociality in divorced samples could be due entirely to the genetic risk transmitted from antisocial parents to their children. Alternatively, environmental disruptions resulting from experiencing divorce could lead to higher rates of child antisociality. Most studies fail to distinguish between genetic and environmental mediation. Many childhood correlates of adult antisocial behavior have been identified. On the within-person level, they include poor school performance and low IQ, and temperamental traits such as hyperactivity and inattention (Bassarath, 2001 and Farrington, 1995), impulsivity, and risk taking behaviors (Bassarath, 2001). Familial correlates include a family history of criminality, poverty, poor parenting (Farrington, 1995), child abuse (Dodge, Bates, & Pettit, 1990), and domestic violence (Jaffee, Moffitt, Caspi, Taylor, & Arseneault, 2002). Childhood behaviors such as affiliation with antisocial peers and substance abuse (Bassarath, 2001) are also correlated with child antisociality. Many of these correlates not only run in families, but tend to co-occur with each other as well. The study of whether these correlates of antisocial behavior, which are often referred to as risk factors, are in fact causal is made more difficult because the correlates are confounded by antisociality in the parents. According to Kraemer et al. (1997) a risk factor should be associated with the specified outcome and precede that outcome. To be considered a causal risk factor, it must be possible to change the factor and manipulation of the factor must result in a change in the specified outcome. Few risk factors for the development of antisociality are likely to pass this stringent causal test. Moreover, while many potential risk factors are present before this outcome unfolds, they are also associated with the presence of a family history of antisocial behavior. As we will show, there is considerable familial resemblance for antisocial behavior, and antisociality has a strong genetic component. While a family history of antisocial behavior is recognized as a strong marker for the development of antisociality, it is often ignored when other markers (e.g., poor parenting, divorce, child abuse) are examined. The possibility of passive gene–environment correlations accounting for the association between these markers and child antisociality is strong. Passive gene–environment correlations are operating when a parent's genes influence the child's environment (e.g., genetically influenced characteristics of the parent lead to engagement in maladaptive parenting of the child). According to Rutter et al. (1997), environmental correlates influenced by the parents’ genes may represent a true environmentally mediated path from parent disorder to child disorder, or the environmental correlate may not have any actual effect, and the transmission instead is genetically mediated. While the concept of gene–environment correlations is not new, many studies continue to ignore the possibility that genetic predispositions in the parents may be leading to the parent's maladaptive behaviors and to genetic predispositions in the children, instead concluding that these maladaptive parental behaviors are causing the child's antisociality. 1.1. Issues addressed The primary aim of this paper is to better understand the etiology of antisocial behavior and the nature and extent of environmental risk. We review potential causes of adult antisocial behaviors, focusing in particular on the factors that appear most likely to be leading to the familial perpetuation of antisociality across generations. We begin by examining evidence of familial resemblance on a range of antisocial behaviors. Then we discuss the relative importance of genetic and shared environmental influences (that is, any aspect of the environment that makes children growing up together more similar for a trait) as well as moderators of these influences. Next we focus on research into potential risk factors in the development of antisocial behavior, highlighting studies that take a family history of antisociality into consideration and examining what is currently known about the potential causal role such correlates might play in the development of antisocial behavior. In reviewing factors influenced by the child's family (such as divorce and poor parenting), we ask whether these environmental phenomena have any association above and beyond their co-occurrence with antisociality in parents (i.e., after controlling for parental antisociality) and whether children already at risk for antisociality (viz. having a family history of antisocial behavior) might be differentially affected by these factors. In conclusion, this paper synthesizes research on the commonly identified potential risk factors for antisociality and evaluates whether these factors remain significantly associated with child antisociality after taking into account genetic propensities shared between parent and child.

Conclusions There is a high degree of familial resemblance for antisocial behaviors and the risk to offspring associated with having an antisocial parent extends to a range of antisocial behaviors. Whether such associations are stronger for boys or girls is as of yet unclear. However, children of both genders are shown to be at increased risk compared to children of non-antisocial parents. A number of correlates of child antisociality have been identified as possible causal risk factors. However, many of these correlates also are associated with parental antisociality, and studies of these correlates traditionally have not taken into account their co-occurrence with a family history of antisocial behavior, thus leaving open the possibility that these correlates are not themselves leading to increased rates of child antisocial behavior, but rather the effects are driven entirely by genetic risk passed from parent to child. Twin and adoption studies show there to be both genetic and environmental influences on antisocial behavior. Additionally, behaviors occurring early in childhood and in adulthood appear to be under greater genetic influence than behaviors first appearing in adolescence. Similarly, relative genetic and environmental influences appear to differ between early-onset, persistent antisocials and adolescent-onset, transitory antisocials. As early-onset antisocial youths are more likely to have a family history of antisociality, this may highlight the importance of considering a family history when evaluating potential risk factors. Seemingly environmental markers such as divorce, poor parenting, and affiliation with deviant peers have been shown to be associated with parental antisocial behavior. While most studies of these variables have failed to take into consideration the presence of gene–environment correlations, a few studies have begun to employ genetically informative designs. Current research suggests that links between divorce and child antisociality are due more to family conflict than to the event of divorce. Additionally, much of this association is further accounted for by high rates of antisociality in the parents in these divorced/high conflict families. However, there is also evidence that parent–child conflict exerts an environmental effect on offspring antisociality. Current research suggests that some of the association between negative parenting and child antisocial behavior is due to heightened rates of parental antisociality in the groups of parents engaging in negative parenting. However, there also appears to be a direct influence of parenting on child antisocial behavior. More extreme forms of maladaptive parenting, specifically child abuse and exposure to domestic violence, have been more extensively investigated, and such studies have employed genetically informative designs. These associations between child antisocial behavior and more extreme forms of negative parental behaviors, such as child abuse, have been found to be both genetically and environmentally mediated. That is to say, genetic propensities shared between parent and child account for some of the association between the parental behavior of child abuse and child antisociality, although there is also an environmental influence independent of shared genes. Even behaviors that may seem at face value to have no association with parental antisocial behavior, such as child affiliation with deviant peers, may be influenced genetically (through active gene–environment correlations) and may be affected by antisociality in the parents. Research into this area is inconclusive, although it appears that parenting and peers may have differential effects for early versus late starters. Overall, children become antisocial through a variety of mechanisms. Genetic propensities passed from parent to child certainly play a role. Additionally, environmental correlates including maladaptive parenting and affiliation with deviant peers are associated with engagement in antisocial behavior and are potential risk factors. However, the size of the effect of these environmental influences is overstated due to ignored gene–environment correlations as shared genetic propensities appear to account for at least a portion of the association between these putatively environmental risk factors and child antisocial behavior. Thus, the true environmental influence is likely smaller than it may initially seem. For divorce/family conflict, it appears that this factor is only associated with child antisociality when the parents are engaging in antisocial behavior themselves. It is important to note, however, that many of these hypotheses have not been tested directly, leaving these conclusions tentative. However, in areas where they have been tested explicitly, with genetically informative designs (e.g., for the more extreme potential risk factors of child abuse and domestic violence), much of the association between the potential risk factor and child antisociality appears to be due to shared genetic propensities between parent and child, leaving a small environmental effect of the risk factor itself. Future research into childhood risk factors for antisociality must take into account genetic propensities towards antisocial behavior shared between parents and children. We suggest four methods for doing this. First, parental antisociality should be assessed and taken into account when other correlates of child antisociality are being evaluated. Second, the possibility of gene–environment interactions where the effect of the environmental variable is stronger when genetic risk is also present should be considered in that some children may be more susceptible to the effect of environmental risk factor in the presence of a parental history of antisocial behavior. Third, as pointed out by Moffitt (1993), distinctions must be made between early onset (life-course persistent) and late onset (adolescent-limited) antisocial individuals as their etiological factors differ. Finally, and most importantly, genetically informative designs should be employed in order to assess the degree to which the role of these factors in the transmission of antisociality from one generation to the next is environmental or genetic in origin. Adoption designs are especially useful in disentangling genetic and environmental effects as they do not have to contend with passive gene–environment correlations since the child's genetics are provided by one set of parents while his/her environment is provided by another set of parents (Rutter, Pickles, Murray, & Eaves, 2001). These designs cannot, however, account for active G–E correlations, and they suffer from the fact that most adoptive families are not high-risk (Rutter et al., 2001). Rutter et al. also suggests the use of blended family designs to study the environmental effects of a biologically unrelated step-parent. Discordant twin studies may also be useful in examining environmental influences that differ between twins, but such investigations will not be useful in studying the effect of family-wide environmental variables. Moffitt, Caspi, and Rutter (2005) suggest three methods to test for environmental mediation: treatment experiments where the environmental variable is randomly assigned, longitudinal studies that include measures of the outcome behavior at multiple times which can help to disentangle whether environmental effects are influencing the child or the child is influencing the environment (see, for example, Burt et al., 2005), and twin and adoption designs (although caution is required when using twin designs to study a measured environment (Purcell & Koenen, 2005)). The Children of Twins design, which focuses on the offspring of identical twins, who, although raised by different sets of parents, are genetically full siblings, and compares them to the offspring of fraternal twins, may be especially useful in studying intergenerational relationships (D’Onofrio et al., 2003). Children of Twins designs allow the disentanglement of genetic and environmental influences on the intergenerational relationship between parent behaviors and child outcomes by comparing child–aunt (or child–uncle) correlations for MZ and DZ twin parents. Genetic factors are implicated when the child–aunt correlation for MZ twins exceeds that of the DZ twins. Equivalent correlations indicate environmental transmission. As each design has its own set of limitations, convergent evidence from multiple research designs will likely be necessary to make more definitive statements about the mechanisms of familial transmission of antisocial behavior.