Journal of Adolescent Research http://jar.sagepub.com

A Child Effects Explanation for the Association Between Family Risk and Involvement in an Antisocial Lifestyle Kevin M. Beaver and John Paul Wright Journal of Adolescent Research 2007; 22; 640 DOI: 10.1177/0743558407306343 The online version of this article can be found at: http://jar.sagepub.com/cgi/content/abstract/22/6/640

Published by: http://www.sagepublications.com

Additional services and information for Journal of Adolescent Research can be found at: Email Alerts: http://jar.sagepub.com/cgi/alerts Subscriptions: http://jar.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations (this article cites 32 articles hosted on the SAGE Journals Online and HighWire Press platforms): http://jar.sagepub.com/cgi/content/refs/22/6/640

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

A Child Effects Explanation for the Association Between Family Risk and Involvement in an Antisocial Lifestyle

Journal of Adolescent Research Volume 22 Number 6 November 2007 640-664 © 2007 Sage Publications 10.1177/0743558407306343 http://jar.sagepub.com hosted at http://online.sagepub.com

Kevin M. Beaver Florida State University

John Paul Wright University of Cincinnati

Most dominant theories of crime and criminality underscore the saliency of the family in the etiology of offending behaviors. Recently, a small pool of research has suggested that elements of the family, especially parents, do not have a lasting impact on children. This line of inquiry argues that once the effects that the child has on the family are taken into account, the relationship between family factors and child outcomes will be reduced substantially. The authors use data from the Cambridge Study in Delinquent Development to test the reciprocal effects between the family and the child. The results of their structural equation models reveal that global measures of family risk have a very limited effect on adolescent involvement in an antisocial lifestyle. However, adolescent embeddedness in an antisocial lifestyle negatively affects family functioning. The authors speak of the implications of their findings. Keywords:

antisocial behavior; child effects; delinquency; family risk; parental socialization

Authors’ Note: The data used in this article were made available by the Inter-University Consortium for Political and Social Research. The data for the Cambridge Study in Delinquent Development, 1961-1981 were originally collected by Dr. David P. Farrington. Neither the collector of the original data nor the consortium bear any responsibility for the analyses or interpretations presented here. We wish to thank the editor and the anonymous reviewers for their insightful comments and suggestions on an earlier draft of this article. Correspondence concerning this article should be addressed to Kevin M. Beaver, Florida State University, College of Criminology and Criminal Justice, 634 West Call Street, Tallahassee, FL 323061127; e-mail: [email protected]. 640 Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Beaver, Wright / Antisocial Lifestyle

641

T

wo explanations have been advanced to account for the voluminous line of research showing that various measures of family life are predictive of a range of child outcomes. First, dominant criminological theories assert that parents and other family factors have a direct effect on their children (Akers, 1998; Gottfredson & Hirschi, 1990; Hagan, 1989; Hirschi, 1969; Sampson & Laub, 1993; Wilson & Herrnstein, 1985). According to typical social science perspectives, “bad” families fail to socialize children effectively and thus are more likely to raise antisocial offspring. “Good” families, in contrast, are more proficient at socializing children and are more likely to raise law-abiding adolescents. The second perspective attempting to explain the association between family life and child outcomes is driven by findings from behavioral genetic research (Cohen, 1999; Harris, 1995, 1998, 2006; Pinker, 2002; Rowe, 1994). This line of literature reveals that once genetic influences are held constant, and once the effects that the child has on the family are taken into account, the relationship between family factors and child outcomes vanishes (Harris, 1995, 1998; Lytton, 1990; Wright & Beaver, 2005). Using data from the Cambridge Study in Delinquent Development, we evaluate these two perspectives. Specifically, we estimate a series of structural equation models (SEMs) to test whether the family directly affects a child’s involvement in an antisocial lifestyle when controlling for reciprocal effects between the child and the family.

The Role of the Family in Criminology One of the most widely accepted “facts” within criminology is that parental and family characteristics have substantive effects on the development of delinquent and criminal behavior in children and adolescents (Gottfredson & Hirschi, 1990; Laub & Sampson, 1988; Loeber & Stouthamer-Loeber, 1986; Patterson, 1982; Rankin & Wells, 1990; Sampson & Laub, 1993; Van Voorhis, Cullen, Mathers, & Garner, 1988; Warr, 1993; Wells & Rankin, 1988). Many of the dominant criminological theories pinpoint certain dimensions of the family as important contributors to delinquency. Social bonding theory, for example, highlights the importance of adolescents becoming attached to their parents to prevent antisocial proclivities from surfacing (Hirschi, 1969). Social learning theory recognizes the importance of antisocial role models, including criminal parents and deviant siblings, in the development of juvenile delinquency (Akers, 1998; Matsueda & Heimer,

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

642

Journal of Adolescent Research

1987). Gottfredson and Hirschi’s (1990) theory identifies parents as the main cause of self-control. Elements of the family life are even incorporated into strain theory, cultural deviance theory, and social disorganization theory, among others (Agnew, 1985, 1992; Hagan, 1989; Patterson, 1982; Tittle, 1995; Wilson & Herrnstein, 1985). Indeed, a close reading of almost every criminological theory, even those with seemingly incompatible viewpoints on the causes of crime, underscores the saliency of family life in the etiology of adolescent delinquency (Agnew, 1985, 1992; Akers, 1998; Elliott, Ageton, & Canter, 1979; Gottfredson & Hirschi, 1990; Hirschi, 1969; Laub & Sampson, 1988; Matsueda & Heimer, 1987; Nye, 1958; Patterson, 1982; Tittle, 1995; Warr, 1993; Wilson & Herrnstein, 1985).

Estimating Family Socialization Effects The most conventional way that social scientists estimate family socialization effects is by using some type of correlational analysis (e.g., ordinary least squares [OLS] regression). In this methodology, a measure of delinquency is regressed on any number of predictor variables (including the family socialization measures). Although some research examines these effects using cross-sectional data, the best designed studies usually use longitudinal research designs that include lagged measures of the independent variables and a lagged measure of the dependent variable to help establish temporal ordering. For example, a self-reported delinquency scale (measured at age 16) could be regressed on family functioning variables measured at an earlier time (at age 13), while controlling for prior delinquent involvement (measured at age 13). If the coefficients for the family functioning scales are significant, then the author would most likely conclude that measures of the family have an effect on adolescent delinquent involvement. A wealth of empirical research using these types of methodologies seems to support the close nexus between parenting practices and adolescent delinquency by showing that measures of parental warmth and/or love, parental supervision, parental discipline, and parent-child involvement are predictive of a wide range of antisocial outcomes (Barnes, Reifman, Farrell, & Dintcheff, 2000; Baumrind, 1991; Burt, Simons, & Simons, 2006; Loeber & StouthamerLoeber, 1986; Patterson, 1982; Sampson & Laub, 1993). In addition, other dimensions of family life, such as sibling delinquency, parental criminality, poverty, low socioeconomic status, single-headed households, and divorce have all been found to be related to maladaptive and sometimes criminal behaviors (Amato, 2001; Farrington, 2003; Gottfredson & Hirschi, 1990; Lipsey & Derzon, 1998; Matsueda & Heimer, 1987; Wilson & Herrnstein, 1985). Taken

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Beaver, Wright / Antisocial Lifestyle

643

together, findings from the family socialization literature appear to reveal that various elements of the family have a powerful effect on shaping and molding children and adolescents (Hartup, 1978; Loeber & StouthamerLoeber, 1986).

Do Parents Matter? Evidence From Behavioral Genetics The relationship between family functioning and child outcomes may seem logical, undeniable, and intuitively obvious. Recent research, however, has called into question whether the family, especially parents, has any lasting effects on their children (Cohen, 1999; Harris, 1995, 1998, 2006; Pinker, 2002; Rowe, 1994; Wright & Beaver, 2005). This line of research does not dispute that there is a correlation between the family and the child’s behaviors and personalities; instead, this research takes exception with the interpretation of the findings. Harris’s (1995, 1998, 2006) group socialization theory perhaps best exemplifies the logic of this argument. She maintains that the relationship between measures of the family and measures of the child can be explained away by two different confounding factors. First, Harris (1995, 1998, 2006) noted that most parental socialization research fails to partial out the effects of genes. According to her theory, the relationship between parental socialization and child outcomes is largely due to the genes that are shared between parent and offspring. For example, take parents who are physically abusive and combative toward their children. Research indicates that abused children have a greater chance of becoming delinquent than children who were not abused (Lewis, Mallouh, & Webb, 1989; Rivera & Widom, 1990; Smith & Thornberry, 1995). The most common interpretation of this association is that the parents’ actions left a lasting and indelible mark on their children that eventually caused the children to engage in some type of criminal act. Harris (1995, 1998, 2006) and other behavioral geneticists (Rowe, 1994), however, have a different explanation. They claim that the genes that caused the parents to be abusive toward their children also may have been inherited by their children. If this is the case, then the genes that caused the parents to act aggressively are the same genes that eventually led the children to become involved in crime and delinquency. In short, the relationship between parenting practices and a child’s behavior and personality traits is spurious and is really due to shared genes (Harris, 1995, 1998, 2006; Rowe, 1994; Wright & Beaver, 2005). A limited amount of empirical work has tested this part of Harris’s (1995, 1998) theory and found support in favor of it. Recently, Wright and Beaver Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

644

Journal of Adolescent Research

(2005) examined whether parents had any effect on the development of childhood self-control. They used data from the Early Childhood Longitudinal Study, Kindergarten Class (ECLS-K) to test this possibility. They first estimated the effects that five different parenting measures had on child levels of self-control without taking genetic factors into account. The results of these models revealed that the parenting measures had a direct and statistically significant effect on self-control in children. However, in subsequent models, when genetic influences were parceled out, the parenting measures dissipated from statistical significance. The findings of this work point to the possibility that Harris’s (1995, 1998) theory has some merit in explaining the relationship between family life and child outcomes. Still, future research needs to address and test Harris’s (1995, 1998, 2006) second reason for questioning the causal role that the family has on the child. In this case, she argues that the causal ordering between measures of parenting and measures of the child’s temperament is reversed. Instead of family experiences affecting the child, the child is actually shaping the functioning of the family. This second explanation is referred to as a “child effects” model (Harris, 1998; Lytton, 1990). Family researchers have documented that parents treat their children very differently depending on how each child behaves (Harris, 1998; Lytton, 1990). A difficult and taxing child, for example, will likely be reprimanded, punished, and disciplined regularly by his or her parents. The child’s sibling, however, who has an easygoing personality and who is relatively obedient, will be much more enjoyable for his or her parents to raise, and punishment will be less frequent. Accordingly, children, depending on their behaviors and personality traits, evoke differential responses from their parents. From a statistical standpoint, most social scientists would interpret the significant association between parenting measures and the child’s behavior as evidence of the role that parents play in the development of problem behaviors. Clearly, this would not necessarily be the case; parenting techniques may simply be used differentially based on the actions and temperament of their children. Behavioral geneticists have long recognized the possibility that children may elicit or evoke certain responses from their environment—a phenomenon referred to as an evocative gene-environment correlation (Moffitt, 2005; Rutter, 2006; Scarr & McCartney, 1983). According to the logic of gene-environment correlations, individual personality traits are highly heritable. These personality traits, in turn, cause people to act in a particular way. For example, an individual with a hair-trigger temper is likely to instigate fights, to cause problems with his or her social relationships, and to wreak havoc in almost every environment he or she encounters. Children, too, will differ tremendously in terms of their Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Beaver, Wright / Antisocial Lifestyle

645

dispositions. These different dispositions will thus affect the way they are parented. A child with a docile temperament will elicit very few negative responses from his or her family, whereas the child’s sibling who is aggressive will be more likely to be punished, spanked, and otherwise disciplined. Thus, each child’s unique genetic predispositions are partially responsible for why his or her parents treat him or her the way that they do. In contrast, bidirectional effects have been integrated into only a very few social science theories, and most, if not all, social science theories have failed to incorporate the logic of gene-environment correlations. Two notable exceptions exist: Patterson’s (1982) coercive family processes theory and Thornberry’s (1987, 1996) theory. However, it should be pointed out that both of these theories focus only on bidirectional effects between the family and the child and ignore gene-environment correlations. According to Patterson’s (1982) theory, adolescents become delinquent because of the dynamic processes that unfold within the household between the child and the parents. Specifically, the child’s problematic behavior brings about a negative response from the parents. In turn, the child responds to his or her parent’s hostile reactions by escalating his or her behavior. Similarly, Thornberry’s (1987, 1996) interactional theory recognizes that weak parentchild attachment may lead to delinquency; however, his perspective also recognizes that involvement in delinquency weakens parent-child attachment. These are classic examples of reciprocal-effects models because the child influences the family, and the family influences the child. It is important to point out that Harris’s view on child effects is strikingly different from Patterson’s (1982) and Thornberry’s (1987, 1996) reciprocaleffects models. The most promising way to determine which of these two perspectives is supported is by directly modeling reciprocal effects between the family and the child.1 According to Harris, if reciprocal effects are modeled directly, then a child would have an effect on the family, but the family would not have an effect on the child (i.e., a unidirectional effect). In stark contrast, both Patterson’s and Thornberry’s theories suggest that if reciprocal effects are examined, then the child would have an effect on the family, but the family would also have an effect on the child (i.e., bidirectional effects). A recent article by Huh, Tristan, Wade, and Stice (2006) sheds some light on whether child effects models or reciprocal-effects models are better able to explain the relationship between parenting practices and child outcomes (see also Chen, Liu, & Li, 2000). In line with the extant literature, they examined the effects that poor parenting had on child behaviors. But, unlike much of the prior parental socialization research, they also examined whether problem behavior elicited poor parenting. To test these two competing perspectives, they used data collected from 496 adolescent girls Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

646

Journal of Adolescent Research

attending four U.S. middle schools. The results of their models revealed strong and consistent effects of children’s behavior on parental management techniques. At the same time, they found little evidence supporting the perspective that poor parenting causes adolescent problem behavior— that is, reciprocal effects between the family and the child were not detected. Huh et al. (2006) concluded that “the high proportion of child effects relative to parent effects seems incompatible with the commonly held view that parenting shapes the behavioral outcomes of children, at least during middle adolescence for females” (p. 196).

The Current Study We build off and extend Huh and his colleagues’ (2006) work in three important ways. First, instead of using an American sample of adolescent girls, we analyze data that contain a sample of boys from England. Our analysis will help to determine whether the findings reported by Huh et al. (2006) are generalizable to male adolescents living outside of the United States. Second, Huh et al. (2006) examined only two parenting measures: parental social support and parental control. In contrast, we develop global measures of family risk that not only include indicators of poor parenting but also measures of the economic well-being of the family. Finally, Huh and colleagues focused on two measures of adolescent problem behaviors: externalizing symptoms and substance abuse. Research reveals, however, that the most serious and chronic adolescent delinquents display a wide range of problems, including delinquency, antisocial personalities, and problematic relationships (Cairns & Cairns, 1994; Gottfredson & Hirschi, 1990; Sampson & Laub, 1993). As a result, we create a measure of the boys’ involvement in an antisocial lifestyle that indexes delinquent involvement, criminogenic characteristics, and various types of social failure. Finally, the crux of our study examines the reciprocal effects between the global measure of family risk and the scale tapping the boys’ antisocial lifestyle.

Method Sample Data for this article come from the well-known Cambridge Study in Delinquent Development, which is a prospective longitudinal study of 411

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Beaver, Wright / Antisocial Lifestyle

647

boys residing in working-class areas of London (Farrington, 2002; West & Farrington, 1977). Initial data collection efforts began in 1961 when the respondents were between the ages of 8 and 9 years old. Since that time, follow-up interviews have been conducted at varying time intervals up through adulthood. One of the attractive qualities of the Cambridge Study in Delinquent Development is the very high retention rate of the original respondents. For example, 95% of the sample was reinterviewed at age 18 (Farrington, 2003). Moreover, information about the boys, their behaviors, their personality traits, and their social life was garnered through multiple reporting sources, including parents, teachers, and peers. Police records were also examined and used to estimate the amount of contact that each respondent had with the criminal justice system. One drawback to the study, however, is that only a subsample of respondents was included for follow-up interviews conducted when the boys were 21 to 22 years old and when the boys were 25 to 26 years old. As a result, the analyses will be limited to three waves of data: one wave collected when the boys were 8 to 9 years old, one wave collected when the boys were 10 to 11 years old, and one wave collected when the boys were 14 to 15 years old. Overall, the data span approximately 6 years of childhood and adolescent development, providing an exceptional opportunity to examine the dynamic relationship between family risk and an adolescent’s involvement in an antisocial lifestyle.

Measures Family Risk Criminologists have long implicated different dimensions of the family as important contributors to the development of delinquent behaviors and antisocial personality traits (Gottfredson & Hirschi, 1990; Patterson, 1982; Sampson & Laub, 1993). Usually, however, researchers break out measures of the family into separate categories (e.g., parents, poverty, and family structure) and examine the independent effects of each measure. There is good reason to believe that families will have the most powerful effect promoting delinquency when multiple criminogenic risk factors “build up” in the same family (Loeber & Stouthamer-Loeber, 1986; Rutter, Quinton, & Hill, 1990). To examine this possibility, we model family risk as a global construct that includes a variety of risk factors cutting across different elements of the family. We developed a measure of family risk for the first two waves of data.

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

648

Journal of Adolescent Research

Measures of Family Risk at Wave 1 (8- to 9-year-olds) Family functioning. Criminological perspectives suggest that childhood and adolescent behavioral patterns are contingent on the quality of parenting (Gottfredson & Hirschi, 1990; Patterson, 1982) and other aspects of family life (Loeber & Stouthamer-Loeber, 1986). To capture the extent to which differences in family life differentially affect adolescents, we developed a measure of Family Functioning. This eight-item scale was created by summing a variety of family variables, such as the discipline quality of parents and the vigilance of parents, with higher scores reflecting more adversities in family functioning (α = .65). A description of all scales and measures can be found in the appendix. Economic disadvantage. Economic disadvantage is another facet of family life often thought to be associated with the development of problem behaviors in adolescence (Lipsey & Derzon, 1998). Measures of low socioeconomic status have been found to be related to conduct problems (Lipsey & Derzon, 1998). In addition, because prior research using the Cambridge Study in Delinquent Development data has found that being raised in a poor family is a risk factor for delinquency (Farrington, 2003), we created a seven-item Economic Disadvantage Scale (α = .68). Measures of Family Risk at Wave 2 (10- to 11-year-olds) Parental supervision. In a comprehensive review of the findings garnered from research analyzing the Cambridge Study in Delinquent Development data, Farrington (2003) concluded that poor parental supervision was one of the strongest independent predictors of adolescent misconduct. As a result, we used a one-item measure capturing the extent to which the parents supervise their child’s behavior. Higher scores on this item reflect more lax parental supervision. Economic disadvantage. We developed a 12-item scale capturing the extent of family economic hardships. Items for this scale tapped the mother’s and father’s education level and whether the family had ever received national assistance (α = .65). Higher scores on the Economic Disadvantage Scale represent more financial deficiencies. Antisocial Lifestyle The focus of our analysis is whether measures of family risk are predictive of boys’ involvement in an antisocial lifestyle. To provide an accurate

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Beaver, Wright / Antisocial Lifestyle

649

and reliable way of operationalizing an antisocial lifestyle, we had to incorporate measures that tapped not only boys’ delinquent behaviors but also their antisocial personalities (Gottfredson & Hirschi, 1990; Wilson & Herrnstein, 1985) and the social problems that are typical of adolescent delinquents (Cairns & Cairns, 1994). Fortunately, the Cambridge Study in Delinquent Development contains a vast array of items that can be used to capture how embedded each boy is in an antisocial lifestyle. Measures of an Antisocial Lifestyle at Wave 1 (8- to 9-year-olds) Parental ratings of the boy’s antisocial lifestyle. A four-item Antisocial Lifestyle Scale was created by using parental reports of a boy’s behaviors and personality traits. Parents were asked to indicate how often the boy was disobedient and acted out (α = .68). Higher scores on this scale reflect more troublesome boys. Peer ratings of the boy’s popularity. At Wave 1, peers were asked to rate each boy’s popularity level. This one-item measure was coded such that higher scores indicate lower popularity levels. Teacher ratings of the boy’s antisocial lifestyle. During Wave 1 interviews, teachers were interviewed to provide information about each boy’s problems that were visible at school. Specifically, teachers responded to questions about the boy’s relationships with other students and how difficult he is to discipline. In total, responses to five items were summed together to form the Teacher Antisocial Lifestyle Scale (α = .56). Measures of an Antisocial Lifestyle at Wave 2 (10- to 11-year-olds) Peer ratings of the boy’s antisocial lifestyle. Three peer nomination measures—each indexing boys’ personality traits and behavior—were collected at Wave 2. These items were designed to measure how daring, how honest, and how troublesome each boy was at school. Responses to these items were added together, with higher scores indicating the presence of more antisocial traits and behaviors (α = .70). Teacher ratings of the boy’s antisocial lifestyle. A five-item Antisocial Lifestyle Scale was created by using teacher ratings of the boys. Unlike the peer ratings scale, the teacher ratings scale assessed the boy’s personality traits, his ability to form and maintain social relationships, and his involvement in misconduct. Items included in this scale ranged from questions

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

650

Journal of Adolescent Research

dealing with the boy’s attention level to how difficult the boy is to discipline (α = .59). Measures of an Antisocial Lifestyle at Wave 3 (14- to 15-year-olds) Criminogenic characteristics. Individual differences in criminogenic personality traits, such as low self-control, are strong and consistent predictors of childhood and adolescent misconduct (Gottfredson & Hirschi, 1990; Pratt & Cullen, 2000; Wilson & Herrnstein, 1985). To capture variability in criminal propensities, we included a seven-item measure of criminogenic characteristics. Teachers were asked to rate the boys on seven different items, including whether they are dare-devils, whether they lack concentration, and whether they daydream (α = .73). Higher scores on this scale indicate more criminogenic traits. Social consequences. The most antisocial children and adolescents often report having unstable peer relationships and exhibit school-related problems (Cairns & Cairns, 1994). To take into account the social struggles faced by delinquents, we included a 10-item scale tapping the social consequences that emanate from deviant behaviors. Teachers were asked, for example, the popularity of the boys, how easily they make friends, and the boys’ school performance (α = .63). Teacher ratings of the boy’s delinquent behavior. At Wave 3 interviews, teachers evaluated the delinquent behaviors of each boy. Teachers were asked to indicate whether the boy lies, whether he is disobedient, whether he is aggressive, and whether he is difficult to discipline. Responses to these items were then summed together to form the Teacher Ratings of Delinquent Behavior Scale (α = .86). Self-reported delinquent behavior. At ages 14 to 15, boys participating in the Cambridge Study in Delinquent Development were asked to complete a Self-Reported Delinquent Behavior Questionnaire. This survey was designed to tap the boys’ involvement in 38 delinquent activities. We summed the responses to each question to form the Self-Reported Delinquent Behavior Scale (α = .90).

Analytical Plan To examine the reciprocal effects between family risk and involvement in an antisocial lifestyle, we use the statistical software package, AMOS, to

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Beaver, Wright / Antisocial Lifestyle

651

estimate a series of SEMs. For these equations, the measure of family risk at Wave 2 is modeled as an unobservable latent construct defined by the Parental Supervision Scale and the Economic Disadvantage Scale. Similarly, the measures of antisocial lifestyle at Waves 2 and 3 are also modeled as a latent variable. The Wave 2 Antisocial Lifestyle Scale is a function of the Peer Ratings Scale and the Teacher Ratings Scale. The Wave 3 Antisocial Lifestyle Scale is defined by the Criminogenic Characteristics Scale, the Social Consequences Scale, the Teacher Reports of Delinquency Scale, and the SelfReported Delinquency Scale. Our first SEM evaluates the measurement of these latent constructs. Importantly, we allowed the error terms to correlate among the four different observable indicators of the Wave 3 Antisocial Lifestyle Scale; no other terms were permitted to covary. We next examine the interrelationships among the Wave 1 and Wave 2 Family Risk Scales and the Wave 1, Wave 2, and Wave 3 Antisocial Lifestyle Scales. To do so, we estimate the stability of family risk over time (i.e., between Waves 1 and 2) and the stability of antisocial lifestyle over time (i.e., from Wave 1 to Wave 2 and from Wave 2 to Wave 3). In addition, the model also examines the longitudinal effect that family risk (at Wave 2) has on involvement in antisocial lifestyle (at Wave 3), when controlling for the Wave 2 antisocial lifestyle measure. To preserve degrees of freedom, the exogenous measures (i.e., the Wave 1 Family Risk Scale and the Wave 1 Antisocial Lifestyle Scale) are modeled as observable indicators and allowed to covary. As a result, the Family Risk Scale (at Wave 1) was created by adding together the Family Functioning Scale and the Economic Disadvantage Scale. Relatedly, the Antisocial Lifestyle Scale was developed by summing the Parental Ratings Scale, the Peer Ratings Scale, and the Teacher Ratings Scale. Our final model is a duplicate of the preceding examples except that it estimates the reciprocal relationship between family risk (at Wave 2) and an antisocial lifestyle (at Wave 2).

Results We begin our analysis by evaluating the measurement of our key constructs: family risk and an antisocial lifestyle. As shown in the top panel of Figure 1, all of the factor loadings for the Family Risk Scale are statistically significant. Similar results are garnered for the measurement of an antisocial lifestyle at Wave 2 and Wave 3. For example, the Peer Ratings Scale and the Teacher Ratings Scale are significant indicators of the Wave 2 antisocial lifestyle factor. For the Wave 3 antisocial lifestyle factor, the

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

652

Journal of Adolescent Research

Criminogenic Characteristics Scale and the Social Consequences Scale have the two strongest factor loadings. The Teacher Ratings of Delinquency Scale and the Self-Reported Delinquency Scale, however, are both significant indicators of the Antisocial Lifestyle Scale. Taken together, the results reported in Figure 1 reveal empirical support for the measurement strategy employed to tap family risk and involvement in an antisocial lifestyle. We next turn to the results of the SEMs that examine the interrelationships between family risk and an antisocial lifestyle during the three waves of data collection. The results shown in Figure 2 reveal four broad findings. First, the model fits the data relatively well (χ2 = 63.64, df = 29, p < .05; minimum sample discrepancy divided by degrees of freedom [CMIN/df] = 2.195, df = 29, p < .05; root mean square error of approximation [RMSEA] = .054). Second, boys’ involvement in an antisocial lifestyle is relatively stable over time. For example, the path coefficient from the Wave 1 antisocial lifestyle measure to the Wave 2 antisocial lifestyle measure is large and statistically significant (β = .61). Likewise, the path coefficient from the Wave 2 antisocial lifestyle measure to the Wave 3 antisocial lifestyle measure is also large and statistically significant (β = .66). Third, and similarly, levels of family risk also remain highly stable from Wave 1 to Wave 2 (β = .91). In other words, those individuals who are being raised in a high-risk family at Wave 1 would, on average, also be being raised in a high-risk family at Wave 2. Fourth, family risk measured at Wave 2 is significantly associated with boys’ score on the antisocial lifestyle measure at Wave 3. This finding is observed even after controlling for prior levels of boys’ antisocial lifestyle. It should be pointed out, however, that the global measure of family risk only has a marginal effect (β = .18) on the measure of antisocial lifestyle. The models calculated up until this point have used a lagged measure of family risk to predict future levels of an antisocial lifestyle. Perhaps a more realistic way to capture the true effect that the family has on the child is by examining the relationship between contemporaneous measures of family risk and an antisocial lifestyle. Specifically, we model the reciprocal effects between family risk and an antisocial lifestyle at Wave 2. Figure 3 presents the results of this model. Once again, the model fits the data moderately well (χ2 = 57.75, df = 27, p < .05; CMIN/df = 2.139, df = 27, p < .05; RMSEA = .053). Of particular interest are the coefficients estimating the reciprocal effects between family risk at Wave 2 and an antisocial lifestyle at Wave 2. As revealed in the middle section of Figure 3, the Family Risk Scale does not have a statistically significant impact on the contemporaneous measure of an antisocial lifestyle. In contrast, however, the Antisocial

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Beaver, Wright / Antisocial Lifestyle

Figure 1 Empirical Assessment of Family Risk and an Antisocial Lifestyle at Waves 2 and 3 Wave 2

Family Risk

.62

.42

Parental Supervision

Economic Disadvantage

Wave 2

Antisocial Lifestyle

.68

.65

Peer Ratings

Teacher Ratings

Wave 3

Antisocial Lifestyle

.46

.76 .70

Criminogenic Characteristics

Social Consequences

.64

Teacher Ratings of Delinquency

Self Reported Delinquency

Note: All factor loadings significant at the .05 level, two-tailed test. Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

653

654

Journal of Adolescent Research

Figure 2 The Longitudinal Relationship Between Family Risk and an Antisocial Lifestyle Wave 1

Wave 2

Wave 3

Family Risk .91* .18* Family Risk

Antisocial Lifestyle

.43*

.66* Antisocial Lifestyle

.61* Antisocial Lifestyle

* Significant at the .05 level, two-tailed test.

Figure 3 Modeling the Reciprocal Effects Between Family Risk and an Antisocial Lifestyle Wave 1

Wave 2

Wave 3

Family Risk .86* .15 Family Risk

.09

Antisocial Lifestyle

.18*

.42*

.66* Antisocial Lifestyle

.56* Antisocial Lifestyle

* Significant at the .05 level, two-tailed test.

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Beaver, Wright / Antisocial Lifestyle

655

Lifestyle factor exerts a statistically significant and positive effect on the measure of family risk. Also, once the reciprocal effects are modeled, the Wave 2 family risk construct does not have a statistically significant effect on the boys’Antisocial Lifestyle Scale measured at Wave 3. Together, these findings indicate the boys in the Cambridge Study in Delinquent Development have an effect on their family’s level of risk but that family risk does not have an effect on the boys’ involvement in an antisocial lifestyle.

Discussion A contentious debate has been waged over the effects that parents and the family have on the development of children (Cohen, 1999; Dodge, 1990; Harris, 1998, 2000, 2006; Lytton, 1990; Pinker, 2002; Rowe, 1994; Vandell, 2000; Wahler, 1990). One line of research reveals that experiences that occur within the household are the driving force for why children turn out the way that they do. Other findings, especially those garnered from behavioral genetic research, indicate that the relationship between family experiences and different child outcomes is much more complex. Advocates of this position maintain that any association between parenting measures and child outcomes is spurious and can be partially accounted for by “child effects” (Harris, 1998; Lytton, 1990). The purpose of this article was to test the merit of these two perspectives. We analyzed data from the Cambridge Study in Delinquent Development to examine the reciprocal effects between family risk and involvement in an antisocial lifestyle. The results of our structural equation models revealed that the measure of family risk did not have an appreciable effect on determining which boys would become embedded in an antisocial lifestyle. Importantly, the lack of an association between family risk and an antisocial lifestyle was observed not only in the longitudinal models but also in cross-sectional analyses that modeled reciprocal effects.2 A very different set of results emerged for the effect that the Antisocial Lifestyle Scale had on family risk. When reciprocal effects between the family and the boy were modeled, the results revealed that boys’ involvement in an antisocial lifestyle significantly increased the risk level of the family.3 Our analysis of the Cambridge Study in Delinquent Development data thus provides support in favor of a childeffects explanation for the relationship between family risk and an antisocial lifestyle (Harris, 1998; Huh et al., 2006; Lytton, 1990). Our findings may seem at odds with a host of social science theories and research, but we offer three potential reasons for why we failed to detect a

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

656

Journal of Adolescent Research

significant effect for the family risk measure. First, much of the parental socialization literature fails to model reciprocal effects between the family and the child (Harris, 1998; Lytton, 1990). Although some of this research attempts to control indirectly for child effects (see Note 1), the methodologies and analytical approaches used in these research designs make it impossible to disentangle the effect of the family on the child from the effect of the child on the family. When more accurate ways of controlling for child effects are used, family and parental influences on child development either evaporate or are attenuated substantially (Huh et al., 2006; Lytton, 1990). Analysis of the Cambridge Study in Delinquent Development data tends to bear this point out. Second, and consistent with prior research, our findings indicated a substantial amount of stability in the antisocial lifestyle measure from childhood to adolescence (Loeber, 1982; Olweus, 1979). These results hint at the likelihood that the origins of antisocial behavior are found very early in life—well before adolescence (Gottfredson & Hirschi, 1990; Karr-Morse & Wiley, 1997). There is good reason to believe, for example, that early life experiences and genetic and/or biological influences interact to predispose people to crime and delinquency (Caspi et al., 2002; Moffitt, 1993). If this is the case, then searching for family contributors to adolescent antisocial behavior may be a futile endeavor. Future research would benefit by exploring the interrelationships between biogenic factors and early childhood environmental influences in the etiology of antisocial behavior. Third, research using genetically arranged data sets that include monozygotic twins and dizygotic twins are able to estimate the proportion of variance in a behavior or trait that is due to the environment and that is due to genetic influences. Specifically, behavioral geneticists have made an important distinction between the shared environment and the nonshared environment. The shared environment is composed of the social factors that are experienced by all siblings within a family. The nonshared environment, in contrast, is made up of the environments and social experiences that are unique to each child. Different peer groups for two siblings exemplify the nature of nonshared environmental influences. Results from a wide range of studies have converged to show that the shared environment has a relatively inconsequential effect on children, but the nonshared environment does have a salient effect (Dunn & Plomin, 1990; Harris, 1998; Reiss, Neiderhiser, Hetherington, & Plomin, 2000; Rowe, 1994; Rutter, 2006). Given that the measure of family risk most likely taps into shared environmental influences, with only limited nonshared environmental effects being measured (e.g., child-specific parental

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Beaver, Wright / Antisocial Lifestyle

657

supervision), perhaps it is not too surprising that we failed to observe a significant effect. We suggest that future criminological research should begin to examine how nonshared environmental influences can affect children differentially. This line of inquiry holds particular promise for identifying the unique features of the environment that lead children from the same household to turn out quite differently, despite their shared familial experiences (Dunn & Plomin, 1990; Harris, 2006). Our findings draw attention to the importance of taking child effects into account when estimating whether family life has an influence on the development of antisocial behaviors (Lytton, 1990). But we would be mistaken if we did not point out the main limitations of our study. First, data from the Cambridge Study in Delinquent Development only include information garnered about male individuals residing in England during the 1960s. This necessarily calls into question whether the findings are generalizable to female individuals and to respondents in other countries. We were unable to test this possibility, but recent work by Huh et al. (2006) reported similar findings using a sample of adolescent girls. Second, our analysis was confined to respondents who were between 8 and 9 years old in the first wave of data. As the SEMs revealed, by the age of 8, involvement in an antisocial lifestyle had already emerged for a substantial number of boys. Admittedly, we would have benefited by extending our models back even further into childhood to determine if measures of family risk had an impact on children very early in life. Third, some of the measures used in the analyses had low reliabilities that may have underestimated the effects of the family risk scales. Finally, we were unable to measure every dimension of family risk. Still, we should point out that the measures comprising the family risk scales have been found to be associated with conduct problems, antisocial behavior, and criminal involvement. Future research, using different data sets and using different research designs, is needed to address these concerns. Nonetheless, studies analyzing the Cambridge Study in Delinquent Development data have provided criminologists with a wealth of information about the causes of crime and delinquency (Farrington, 2003). Our research adds to this literature by revealing that families, and parents within families, may not necessarily be facilitating the development of antisocial conduct but instead are simply absorbing the problems associated with raising a chronically delinquent child. To pretend that children are passive receptors of their environment and do not actively sculpt their families is wrong. A more complete and accurate explanation of behavior—one that recognizes the importance of child effects—is needed to understand more fully the causes of crime and delinquency.

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

658

Journal of Adolescent Research

Appendix Description of Variables and Scales Family risk measures at Wave 1 (ages 8 to 9) Family Functioning

Economic Disadvantage

Scale created by summing the following eight items reported on by the parent and the interviewer (α = .65): 1. Discipline quality of mother 2. Maternal attitude 3. Physical neglect of boy 4. Praise by parents 5. Paternal attitude 6. Vigilance of parents 7. Discipline quality of father 8. Inconsistency in disagreement between parents Scale created by summing the following seven items (α = .68): 1. Family size 2. Housing of family 3. Housing—care of interior 4. Income of family 5. Social agencies for family 6. Broken home before age 10 7. Socioeconomic status of family

Family risk measures at Wave 2 (ages 10 to 11) Parental Supervision Economic Disadvantage

A one-item measure determining the degree of supervision the parents give their child Scale created by summing and standardizing the following 12 items (α = .65): 1. Whether or not a number of different social agencies had contact with the family 2. Whether or not the family had ever received national assistance 3. Socioeconomic status of the father 4. Socioeconomic status of the mother 5. Education of father 6. Education of mother 7. Age of building 8. Care of interior 9. Dilapidation of premises 10. Dwelling has a fixed bath with running water 11. Whether or not the family had moved one or more times 12. Whether or not the boy shares a bed with others

Antisocial lifestyle measures at Wave 1 (ages 8 to 9) Parental Ratings of the Boy’s Antisocial Lifestyle

Scale created by summing the following four items (α = .68): 1. Parent rating: acting out 2. Parent rating: adventurousness of boy

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Beaver, Wright / Antisocial Lifestyle

659

Appendix (continued)

Peer Rating of Boys’ Popularity Teacher Ratings of the Boy’s Antisocial Lifestyle

3. Parent rating: conduct disorder of boy 4. Parent rating: obedience of boy A one-item question measuring boys’ popularity as reported by peer nominations Scale created by summing the following four items (α = .56): 1. Teacher rating: lazy boy 2. Teacher rating: doesn’t care 3. Teacher rating: difficult relations 4. Teacher rating: difficult to discipline

Antisocial lifestyle measures at Wave 2 (ages 10 to 11) Peer Ratings of Boys’ Antisocial Lifestyle

Teacher Ratings of the Boy’s Antisocial Lifestyle

Scale created by summing the following three items (α = .70): 1. Peer rating: daring 2. Peer rating: honest 3. Peer rating: troublesome Scale created by summing the following four items (α = .59): 1. Teacher rating: lazy boy 2. Teacher rating: does not care 3. Teacher rating: difficult relations 4. Teacher rating: difficult to discipline

Antisocial lifestyle measures at Wave 3 (ages 14 to 15) Criminogenic Characteristics

Social Consequences

Scale created by summing the following seven items (α = .73): 1. Teacher rating: lazy 2. Teacher rating: lacks concentration 3. Teacher rating: restless 4. Teacher rating: daydreams 5. Teacher rating: attention seeking 6. Teacher rating: dare-devil 7. Teacher rating: reaction to criticism or punishment Scale created by summing the following eight items (α = .63): 1. Teacher rating: position in class 2. Teacher rating: reading ability 3. Teacher rating: popular 4. Teacher rating: makes friends 5. Parental rating: conduct disorder of boy 6. Parental rating: parental approval of boy 7. Teacher rating: school performance of boy 8. Teacher rating: punctuality (continued)

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

660

Journal of Adolescent Research

Appendix (continued) Teacher Ratings of the Boy’s Delinquency

Self-Reported Delinquency

Scale created by summing the following four items (α = .86): 1. Teacher rating: general aggressiveness 2. Teacher rating: disobedient 3. Teacher rating: difficult to discipline 4. Teacher rating: lies Scale created by summing the following thirty eight items (α = .90): 1. Riding a bicycle without lights after dark 2. Driving under the age of 16 3. Gang activity 4. Playing truant from school 5. Deliberately travel without ticket or paying wrong fare 6. Letting off fireworks in the street 7. Taking money from home, no intention of returning it 8. Stealing car or motor bike for joyriding 9. Smashing, slashing, or damaging things in public places 10. Annoying, insulting, and/or fighting people or strangers in street 11. Breaking into big store, garage, warehouse, pavilion, etc. 12. Breaking into a small shop without necessarily stealing 13. Stealing things out of cars 14. Carrying some kind of weapon like a knife or a cosh 15. Attacking enemy/rival gang (without weapon) in public 16. Breaking the windows of empty houses 17. Using weapon in a fight—knife, cosh, razor, bottle, etc. 18. Drinking alcoholic drinks in pubs under the age of 18 19. Going into pub bars under the age of 16 20. Stealing from big stores, supermarkets while open 21. Stealing things from small shops or tradesmen while open 22. Deliberately littering street by smashing bottles etc. 23. Buying or accepting stolen goods 24. Accomplishing premeditated burglary from a house 25. Accomplishing unpremeditated burglary from a house 26. Stealing pedal cycle and keeping it 27. Struggling or fighting to get away from a policeman 28. Attacking a policeman trying to arrest someone else 29. Stealing school property worth more than about 5P 30. Stealing from employers (more than 50P) while working 31. Trespassing railway lines, good yards, empty houses 32. Going to X films underage 33. Gambling (more than 1 pound a week) under age 16 34. Regularly smoking cigarettes under the age of 15 35. Stealing goods and/or money from slot machines and/or juke boxes etc. 36. Stealing from people’s clothes hanging up anywhere 37. Obtaining money by false pretenses 38. Taking illegal drugs like purple hearts or smoking reefers

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Beaver, Wright / Antisocial Lifestyle

661

Notes 1. Some researchers have sought to model child effects by introducing a prior measure of the dependent variable as a control variable. For example, if a delinquency scale at Wave 3 is used as the dependent variable, a delinquency scale at Wave 2 has been introduced to the statistical models as a way to help rule out child effects. This, however, does not work. Including a prior measure of delinquency only changes what is being predicted. Instead of predicting scores on the delinquency scale, now the statistical models are predicting change in delinquency between Wave 2 and Wave 3; this does not help rule out the possibility of child effects. Even so, some researchers have also assumed erroneously that controlling for prior levels of the independent and dependent variables will help to model child effects. In this case, structural equation models (SEMs) or some other types of statistical models are calculated whereby parenting measures (at Wave 3) are used to predict scores on the Delinquency Scale (at Wave 3). To control for the stability of parenting behaviors across waves, Wave 2 parenting measures are used to predict Wave 3 parenting measures. The same procedure is used for the Delinquency Scale: Wave 2 delinquency is used to predict involvement in delinquency at Wave 3. The measures at Wave 3 now reflect change in parenting scores (between Waves 2 and 3) and change in delinquency (between Waves 2 and 3). Next, the Wave 3 parenting measures (that measure change in parenting) are used to predict Wave 3 delinquency scores (that measure change in delinquency). If a statistically significant effect is found, then researchers proclaim that the evidence suggests that changes in parenting causes changes in their child’s behavior. Again, however, this is incorrect. A child-effects explanation can account for this association just as effortlessly as parental socialization perspectives. According to the logic of child effects, a change in the child’s behavior can bring about a change in the parent’s behavior. Child effects are estimated most accurately by using SEMs that model the reciprocal effects between contemporaneous measures of the family and of the child. 2. The lagged effect of family risk on an antisocial lifestyle was marginally significant in the longitudinal model (Figure 2). However, once reciprocal effects were modeled, the family risk effect was reduced to statistical insignificance. 3. At first glance, it might seem counterintuitive to argue that a child could have an effect on the functioning of the family, especially the economic conditions of the family. However, this is an empirical question and, as our results revealed, children and adolescents who are heavily embedded within an antisocial lifestyle may actually have some influence on the functioning of the family, including the economic well-being. How could this be the case? We offer one possible explanation. Imagine a child who is in constant trouble at school, who causes problems at home, and who is always in trouble with the criminal justice system. The parent or parents may have to leave their place of employment to remove their child from school or from the police station. The child, therefore, is interfering with the parents’ ability to function at work. In addition, the child may also need psychiatric help or may need a lawyer to represent him or her. From this point of view, it takes no stretch of the imagination to see that a behaviorally disordered child may cause economic problems for the entire family. Future qualitative and quantitative research is needed to explore the potential ways in which children and adolescents can have deleterious effects on different dimensions of the family.

References Agnew, R. (1985). A revised strain theory of delinquency. Social Forces, 64, 151-167. Agnew, R. (1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47-87.

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

662

Journal of Adolescent Research

Akers, R. L. (1998). Social learning and social structure: A general theory of crime and deviance. Boston: Northeastern University Press. Amato, P. R. (2001). Children of divorce in the 1990s: An update of the Amato and Keith (1991) meta-analysis. Journal of Family Psychology, 15, 355-370. Barnes, G. M., Reifman, A. S., Farrell, M. P., & Dintcheff, B. A. (2000). The effects of parenting on the development of adolescent alcohol misuse: A six-wave latent growth model. Journal of Marriage and the Family, 62, 175-186. Baumrind, D. (1991). The influence of parenting style on adolescent competence and substance use. Journal of Early Adolescence, 11, 56-95. Burt, C. H., Simons, R. L., & Simons, L. G. (2006). A longitudinal test of the effects of parenting and the stability of self-control: Negative evidence for the general theory of crime. Criminology, 44, 353-392. Cairns, R. B., & Cairns, B. D. (1994). Lifelines and risks: Pathways of youth in our time. Cambridge, UK: Cambridge University Press. Caspi, A., McClay, J., Moffitt, T. E., Mill, J., Martin, J., Craig, I. W., et al. (2002). Role of genotype in the cycle of violence in maltreated children. Science, 297, 851-854. Chen, X., Liu, M., & Li, D. (2000). Parental warmth, control, and indulgence and their relations to adjustment in Chinese children. Journal of Family Psychology, 14, 401-419. Cohen, D. B. (1999). Strangers in the nest: Do parents really shape their child’s personality, intelligence, or character? New York: John Wiley. Dodge, K. A. (1990). Nature versus nurture in childhood conduct disorder: It is time to ask a different question. Developmental Psychology, 26, 698-701. Dunn, J., & Plomin, R. (1990). Separate lives: Why siblings are so different. New York: Basic Books. Elliott, D. S., Ageton, S. S., & Canter, R. J. (1979). An integrated theoretical perspective on delinquent behavior. Journal of Research in Crime and Delinquency, 16, 3-27. Farrington, D. P. (2002). Cambridge study in delinquent development, 1961-1981 [computer file]. Ann Arbor, MI: Inter-University Consortium for Political and Social Research. Farrington, D. P. (2003). Key results from the first forty years of the Cambridge Study in Delinquent Development. In T. P. Thornberry & M. D. Krohn (Eds.), Taking stock of delinquency: An overview of findings from contemporary longitudinal studies (pp. 137-183). New York: Kluwer. Gottfredson, M., & Hirschi, T. (1990). A general theory of crime. Stanford, CA: Stanford University Press. Hagan, J. (1989). Structural criminology. New Brunswick, NJ: Rutgers University Press. Harris, J. R. (1995). Where is the child’s environment? A group socialization theory of development. Psychological Review, 102, 458-489. Harris, J. R. (1998). The nurture assumption: Why children turn out the way they do. New York: The Free Press. Harris, J. R. (2000). Socialization, personality development, and the child’s environments: Comment on Vandell (2000). Developmental Psychology, 36, 711-723. Harris, J. R. (2006). No two alike: Human nature and human individuality. New York: Norton. Hartup, W. W. (1978). Perspectives on child and family interaction: Past, present, and future. In R. M. Lerner & G. B. Spanier (Eds.), Child influences on marital and family interaction: A life-span perspective (pp. 23-46). San Francisco: Academic Press. Hirschi, T. (1969). Causes of delinquency. Berkeley: University of California Press. Huh, D., Tristan, J., Wade, E., & Stice, E. (2006). Does problem behavior elicit poor parenting? A prospective study of adolescent girls. Journal of Adolescent Research, 21, 185-204.

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Beaver, Wright / Antisocial Lifestyle

663

Karr-Morse, R., & Wiley, M. S. (1997). Ghosts from the nursery: Tracing the roots of violence. New York: The Atlantic Monthly Press. Laub, J. H., & Sampson, R. J. (1988). Unraveling families and delinquency: A reanalysis of the Gluecks’ data. Criminology, 26, 355-380. Lewis, D. O., Mallouh, C., & Webb, J. (1989). Child abuse, delinquency, and violent criminality. In D. Cicchetti & V. Carlson (Eds.), Child maltreatment: Theory and research on the causes and consequences of child abuse and neglect (pp. 707-721). New York: Cambridge University Press. Lipsey, M. W., & Derzon, J. H. (1998). Predictors of violent or serious delinquency in adolescence and early adulthood: A synthesis of longitudinal research. In R. Loeber & D. P. Farrington (Eds.), Serious and violent juvenile offenders: Risk factors and successful interventions (pp. 86-105). Thousand Oaks, CA: Sage. Loeber, R. (1982). The stability of antisocial and delinquent child behavior: A review. Child Development, 53, 1431-1446. Loeber, R., & Stouthamer-Loeber, M. (1986). Family factors as correlates and predictors of juvenile conduct problems and delinquency. In M. Tonry & N. Morris (Eds.), Crime and justice: A review of research (Vol. 7, pp. 29-149). Chicago: University of Chicago Press. Lytton, H. (1990). Child and parent effects in boys’ conduct disorder: A reinterpretation. Developmental Psychology, 26, 683-697. Matsueda, R. L., & Heimer, K. (1987). Race, family structure, and delinquency: A test of differential association and social control theories. American Sociological Review, 52, 826-840. Moffitt, T. E. (1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674-701. Moffitt, T. E. (2005). The new look of behavioral genetics in developmental psychopathology: Gene-environment interplay in antisocial behaviors. Psychological Bulletin, 131, 533-554. Nye, I. F. (1958). Family relationships and delinquent behavior. New York: John Wiley. Olweus, D. (1979). Stability of aggressive reaction patterns in males: A review. Psychological Bulletin, 86, 852-875. Patterson, G. R. (1982). Coercive family process. Eugene, OR: Castilia. Pinker, S. (2002). The blank slate: The modern denial of human nature. New York: Viking. Pratt, T. C., & Cullen, F. T. (2000). The empirical status of Gottfredson and Hirschi’s general theory of crime: A meta-analysis. Criminology, 38, 931-964. Rankin, J. H., & Wells, L. E. (1990). The effect of parental attachments and direct controls on delinquency. Journal of Research in Crime and Delinquency, 27, 140-465. Reiss, D., Neiderhiser, J. M., Hetherington, E. M., & Plomin, R. (2000). The relationship code: Deciphering genetic and social influences on adolescent development. Cambridge, MA: Harvard University Press. Rivera, B., & Widom, C. S. (1990). Childhood victimization and violent offending. Violence and Victims, 5, 19-35. Rowe, D. C. (1994). The limits of family influences: Genes, experience, and behavior. New York: Guilford. Rutter, M. (2006). Genes and behavior: Nature-nurture interplay explained. Malden, MA: Blackwell. Rutter, M., Quinton, D., & Hill, J. (1990). Adult outcomes of institution-reared children. Males and females compared. In L. N. Robins & M. Rutter (Eds.), Straight and devious pathways from childhood to adulthood (pp. 135-157). Cambridge, UK: Cambridge University Press. Sampson, R. J., & Laub, J. H. (1993). Crime in the making: Pathways and turning points through life. Cambridge, MA: Harvard University Press.

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

664

Journal of Adolescent Research

Scarr, S., & McCartney, K. (1983). How people make their own environments: A theory of genotype → environment effects. Child Development, 54, 424-435. Smith, C., & Thornberry, T. P. (1995). The relationship between childhood maltreatment and adolescent involvement in delinquency. Criminology, 33, 451-477. Thornberry, T. P. (1987). Toward an interactional theory of delinquency. Criminology, 25, 863-891. Thornberry, T. P. (1996). Empirical support for interactional theory: A review of the literature. In J. D. Hawkins (Ed.), Delinquency and crime: Current theories (pp. 198-235). New York: Cambridge University Press. Tittle, C. R. (1995). Control balance: Toward a general theory of deviance. Boulder, CO: Westview. Vandell, D. L. (2000). Parents, peer groups, and other socializing influences. Developmental Psychology, 26, 699-710. Van Voorhis, P., Cullen, F. T., Mathers, R. A., & Garner, C. C. (1988). The impact of family structure and quality on delinquency: A comparative assessment of structural and functional factors. Criminology, 26, 235-261. Wahler, R. G. (1990). Who is driving the interactions? A commentary on “child and parent effects in boys’ conduct disorder.” Developmental Psychology, 26, 702-704. Warr, M. (1993). Parents, peers, and delinquency. Criminology, 31, 17-40. Wells, L. E., & Rankin, J. H. (1988). Direct parental controls and delinquency. Criminology, 26, 263-285. West, D. J., & Farrington, D. P. (1977). The delinquent way of life. New York: Crane Russak. Wilson, J. Q., & Herrnstein, R. J. (1985). Crime and human nature: The definitive study of the causes of crime. New York: Free Press. Wright, J. P., & Beaver, K. M. (2005). Do parents matter in creating self-control in their children? A genetically informed test of Gottfredson and Hirschi’s theory of low self-control. Criminology, 43, 1169-1202.

Kevin M. Beaver is an assistant professor in the College of Criminology and Criminal Justice at Florida State University. He received his doctoral degree and his master’s degree in criminal justice from the University of Cincinnati. His current research focuses on examining the biosocial and genetic correlates to serious violence. John Paul Wright is an associate professor of criminal justice in the Division of Criminal Justice at the University of Cincinnati. He has published widely on topics that include the development of criminal offending across the life course, the origins of self-control, and biosocial criminology.

Downloaded from http://jar.sagepub.com at FLORIDA STATE UNIV LIBRARY on October 11, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.