Self-control

Self-control Order Description Please answer below questions using article . "Discuss the evidence pertaining to the biological/genetic contributors to self-control. Based on this evidence, do you think that parental socialization is the only cause of variation in self-control? " CRIMINOLOGY VOLUME 43 NUMBER 4 2005 1169 DO PARENTS MATTER IN CREATING SELFCONTROL IN THEIR CHILDREN? A GENETICALLY INFORMED TEST OF GOTTFREDSON AND HIRSCHI’S THEORY OF LOW SELF-CONTROL* JOHN PAUL WRIGHT University of Cincinnati KEVIN M. BEAVER University of Cincinnati KEYWORDS: ADHD, behavior patterns, genetics, parental influence, self-control Gottfredson and Hirschi’s general theory of crime (1990) has generated an abundance of research testing the proposition that low self-control is the main cause of crime and analogous behaviors. Less empirical work, however, has examined the factors that give rise to low self-control. Gottfredson and Hirschi suggest that parents are the sole contributors for either fostering or thwarting low self-control in their children, explicitly discounting the possibility that genetics may play a key role. Yet genetic research has shown that ADHD and other deficits in the frontostriatal system are highly heritable. Our research thus tests whether “parents matter” in creating low self-control once genetic influences are taken into account. Using a sample of twin children we find that parenting measures have a weak and inconsistent effect. We address the conceptual and methodological issues associated with the failure to address genetic influences in parenting studies. More than a decade ago, Gottfredson and Hirschi (1990) set forth a general theory of crime that assigned low self-control as the causal factor in the etiology of crime and numerous analogous behaviors. Since that * We would like to thank the editor for his commitment to scholarly discourse. Direct all correspondence to John Paul Wright, Division of Criminal Justice,. 600 Dyer Hall, Ml 0389,University of Cincinnati, Cincinnati, OH 45221 or email at [email protected]. 1170 WRIGHT AND BEAVER time, the theory has occupied a fundamental position in criminology and has generated an abundance of empirical tests. Its robustness is evident in the recent meta-analytic review by Pratt and Cullen (2000), who found that, across various samples and measurement techniques, low self-control is a salient predictor of criminal behavior. With such support, it should be no surprise that low self-control theory continues to be at the heart of much criminological debate and investigation (Geis, 2000; Marcus, 2004; Sampson and Laub, 1995). Gottfredson and Hirschi’s theory focused on the association between low self-control and crime. Indeed, it is this association that has generated the greatest amount of empirical interest. Their hypotheses that link the development of self-control in children to parental behaviors, however, have been less frequently considered. Parents, they maintain, play the decisive role in either fostering or thwarting the development of low selfcontrol. Borrowing heavily from the work of Patterson (1982), Gottfredson and Hirschi assert that parents who effectively monitor and supervise their children, and who recognize and respond to their child’s antisocial behavior will effectively instill self-control in that child. Parents who fail to engage in such parental management techniques will subsequently fail to help their children develop the ability to resist situational temptations. Given the overwhelming support linking low self-control to crime and analogous behaviors (Pratt and Cullen, 2000), the paucity of research examining the factors that give rise to low self-control is somewhat surprising. We have, for instance, been able to locate only a handful of empirical studies that test them. The findings, in general, have been favorable to Gottfredson and Hirschi’s position that effective child-rearing practices are predictive of self-control in children (but see Cochran, Wood, Sellers, Wilkerson, and Chamlin, 1998). Even so, the quality of that evidence is circumspect, for reasons we will detail later. Gottfredson and Hirschi attribute the development of low self-control in children solely to parenting practices—rejecting outright potential genetic effects. At the same time, a large and growing body of clinical and behavioral genetic research has found that impulsivity, attention deficit hyperactivity disorder (ADHD), and hyperactivity—concepts related closely to Gottfredson and Hirschi’s construct—are highly heritable (Price, Simonoff, Waldman, Asherson and Plomin, 2001; Rietveld, Hudziak, Bartels, van Beijsterveldt and Boomsma, 2003). In a review of the genetic research on ADHD, for instance, Spencer and his colleagues concluded that “the mean heritability of ADHD... is approximately 0.75, which means that about 75% of the etiological contribution to this disorder is genetic” (2002:6). Indeed, after a review of the evidence, the National Institute of Mental Health published a widely cited brochure DO PARENTS MATTER IN SELF-CONTROL? 1171 stating that scientists “are finding more and more evidence that ADHD does not stem from home environment, but from biological causes” (2003:13; for a meta-analytic review of the association between ADHD and criminal behavior see Pratt, Cullen, Blevins, Daigle and Unnever, 2002). Other authors have reached similar conclusions. Barkley (2005), for example, notes that when DSM criteria are used to measure ADHD that studies indicate it to be 97 percent heritable. “This trait,” he argues, “is more inherited than any dimension of human personality” (14). The potential for genetic heritability to influence levels of low selfcontrol in children poses a serious counter argument to Gottfredson and Hirschi’s parenting thesis. For instance, various scholars now openly question whether “parents matter” in the development of their children’s personalities. Harris (1998), for example, has argued forcibly that the effects of parenting on child outcomes have been overstated, and that in most instances, parents “don’t matter” when it comes to the child’s personality (see also Cohen, 1999; Wright and Cullen, 2001). Citing evidence from behavioral genetic studies (see for example, O’Connor, Neiderhiser, Reiss, Hetherington and Plomin, 1998; Plomin, 1995; Scarr and McCartney, 1983), Harris argues that parental socialization practices are likely to be inconsequential once individual differences in parent and child temperament and genetic heritability are accounted for (Cohen, 1999; Pinker, 2002). Harris’s position is more complex than the simple statement that “parents don’t matter” implies. Her critique of the parenting research in general raises serious questions about the validity of many social science findings relating parenting practices to offspring conduct. The vast majority of research on parenting, she notes, typically employs samples that measure one child and one parent, usually the mother, under the assumption that inferences can be made to other children in the household (Rowe, 1994; Walsh, 2002). More detailed research has revealed, however, that parents enjoy differential relationships with their children. They may treat one with hostility, yet pamper another (Harris, 1998). Children too, when asked, often report substantial differences in their perceptions of their parents (Reiss, Neiderhiser, Hetherington and Plomin, 2000). According to Harris, numerous “micro-environments” exist within any home. It is these micro-environments, or child-specific parenting behaviors as opposed to measures of global parenting, that likely differentiate children. Such micro-environments are typically not examined with standard social methodologies (SSM) (see also Walsh, 2002). Nor do SSMs account for commonalities due to genetic similarities. Findings from a broad array of studies converge to show that many temperamental factors are highly heritable (Caspi, Roberts and Shiner, 2005). In turn, associations between parents’ behaviors and the behaviors 1172 WRIGHT AND BEAVER of their children may be confused for “statistically significant parenting effects” in study designs that are insensitive to biological similarities between subjects within the home. For example, mothers who are hostile and cold are more likely to be emotionally removed from their daughters’ lives. Their daughters, in turn, are more likely to be hostile and cold. SSMs would correlate the daughters’ hostility with the mother’s removed parenting style and likely infer that maternal hostility caused daughter hostility—all without any recognition that the two phenotypes share common genetic backgrounds. These limitations, Harris argues, likely misspecify or overstate the effects parents may have on their children’s traits and behaviors.1 We should also add that numerous behavioral genetic studies have failed to detect significant shared environmental effects (Dunn and Plomin, 1990; Neiderhiser, Reiss, Hetherington and Plomin, 1999; Plomin, Owen and McGuffin, 1994). Our purpose here is twofold. First, we examine the effects of parenting on levels of self-control in kindergarten and first-grade children. We do so with a national dataset that contains mother and teacher reports of child self-control. As do other SSM studies, we use OLS models with appropriate controls for demographic and neighborhood influences. Second, and more important, we also use a sample of twins, taken from the same dataset, to assess the influence of parenting factors. The use of a twin dataset allows us to account for the shared genetic variance between twins. Moreover, we use hierarchical linear regression (HLM) analyses to control for the clustering of observations caused by genetic similarities. We contrast results garnered through traditional social science methodologies with those generated from the sample of twins. It appears, to foreshadow our results somewhat, that Harris’s critiques should no longer be ignored. EFFECTS OF PARENTING ON LOW SELF-CONTROL A long line of literature has placed parents at the forefront of criminological theories and research (Loeber and Stouthamer-Loeber, 1986; Patterson, 1982). Such research has tended to focus on the ways various parenting styles, usually measured as global parenting constructs, shape children’s behavioral patterns. Given the dominant role of parents in criminology, it is interesting that very little research has been conducted on how parents influence their children’s self-control—especially given the empirical attention dedicated to the theory. 1. The interpretation of parenting correlations with child behavioral outcomes is made difficult by at least three factors: First, parents and their children share genes. Second, external biological factors, such as neurotoxins, may influence both parent and child behavior. Third, temporal ordering is very difficult to establish. Child traits and behaviors likely influence parenting behaviors and vice versa. DO PARENTS MATTER IN SELF-CONTROL? 1173 Using data from the Cambridge Study in Delinquent Development, Polakowski (1994) examined the role of parental supervision on child selfcontrol. His analysis employed two measures of child-rearing practices— parent’s watchfulness and parental supervision—both garnered from interviews with social workers. These variables were used as indicators for a latent measure of supervision. Although not a complete test of Gottfredson and Hirschi’s proposition, Polakowski’s findings, generated from structural equation models (SEM), were generally consistent with Gottfredson and Hirschi’s hypothesis. Children whose parents were vigilant, had, on average, more self-control (see also Lynskey, Winfree, Esbensen and Clason, 2000; Pratt, Turner and Piquero, 2004). Analyzing data from sixth-grade male students, Feldman and Weinberger (1994), explored the relationships among parenting practices, delinquent behavior and childhood self-control, or what they termed “selfrestraint.” Their results indicated that parental management was positively associated with higher levels of child self-restraint, yet did not have a direct effect on child misbehavior. Similarly, two additional studies, conducted by Gibbs and his associates, examined retrospective accounts of parental management practices on levels of self-control in a sample of college students. The first—Gibbs, Giever and Martin (1998)—found tentative support for the role parents play in fostering low self-control. Gibbs et al. measured forty characteristics of parental management styles and another forty of low self-control. Through a series of path diagrams they found that parental management had a significant and direct effect on low self-control (Beta=.28). Likewise, Gibbs, Giever and Higgins (2003) performed another analysis on a sample of college students, again using SEMs. Their findings paralleled those reported in their 1998 study. Parental management practices maintained a positive relationship with low selfcontrol, with the coefficient being moderate in magnitude (Beta=.26). Similar results were garnered in a study replicated by Higgins (2002). Hay (2001) also examined the effects of parenting on low self-control in a sample of 197 urban high school students. Their analysis also included the two parenting measures—monitoring and discipline—along with a selfreport measure of low self-control in his analyses. The results provided partial support in favor of Gottfredson and Hirschi’s theory. Hay’s analyses revealed that parental monitoring, but not discipline, was significantly associated with child low self-control, even after controls were introduced for early childhood antisocial behavior. Hay also analyzed an alternative model that combined the two parenting scales into one monitoring-discipline measure. This was significantly and inversely related to low self-control (Beta=-.24). 1174 WRIGHT AND BEAVER More recently, Unnever, Cullen, and Pratt (2003) found evidence linking parenting practices to offspring low self-control. Data for their study came from 2,437 middle school students in Virginia. Similar to Hay (2001), Unnever and his associates employed measures of parental monitoring and of consistent punishment. Their analysis also included Grasmick, Tittle, Bursik and Arneklev’s (1993) low self-control scale. Their findings indicated that monitoring and consistent punishment were significantly related to low self-control, even when controlling for the child’s level of ADHD.2 These studies suggest that Gottfredson and Hirschi’s theory on the development of low self-control (1990) is at least partly correct. Under the assumptions of SSMs, parenting practices appear to have some influence on offspring low self-control. The strength of the relationship between parenting practices and child self-control, however, appears to be moderate at best. More important, as we will show, there is reason to cast doubt over the validity of this body of research. GENETIC CONTRIBUTIONS TO LOW SELF-CONTROL Gottfredson and Hirschi explicitly discount the possibility that low selfcontrol may have a genetic component: “the magnitude of the ‘genetic effect’” they say, “is near zero” (60). A large body of literature, however, has arrived at a very different conclusion. In a recent study that examined heritability of attention problems in twins drawn from the Netherlands Twin Registry, Rietveld and colleagues (2002) found that heritability estimates varied between .68 and .76, depending on the age of the twins. Their longitudinal study of 3,853 twin pairs also found that shared environment, which is typically conceived of as family environment, had little effect on the child’s level of overactivity. Similar to Barkley’s (1997) conclusions, Rietveld and colleagues suggest strongly that attention problems are due more to genetic factors than to environmental influences. In a similar vein, Mick, Biederman, Prince, Fischer and Faraone (2002) found the strongest risk factor for childhood ADHD was having a parent with ADHD. Children with an ADHD parent were eight times more likely to be diagnosed with ADHD. The effects of parental ADHD were stronger than fetal exposure to drugs, alcohol, and cigarette smoke, than having a low birth weight, and than being born into an economically disadvantaged family. These findings suggest that even when common 2. Some studies do attempt to control partially for child effects, such as Hay (2001) and Unnever et al. (2003), by using autoregressive statistical models. Others do not rely on parental reports and instead utilize individual recollections of parenting behaviors (Gibbs et al., 1998). DO PARENTS MATTER IN SELF-CONTROL? 1175 environmental risk factors related to offending are taken into account, genetic factors continue to exert a substantial effect on the child (Reiss, Neiderhiser, Hetherington and Plomin, 2000). Attention problems, problems with hyperactivity, and problems with impulsivity have repeatedly been shown to have a substantial genetic component. Given the close correspondence between Gottfredson and Hirschi’s conception of low self-control and the diagnostic criteria for ADHD, it is likely that low self-control is also highly heritable. Indeed, in the vernacular of developmental neuropsychologists, “executive controls” are composed of the ability to regulate emotions, to control impulses, to focus appropriately on the task at hand, and to delay gratification. These capacities are housed in the frontal, orbital-frontal and prefrontal cortex of the brain, which are part of the larger frontostriatal system (Aron, Robbins and Poldrack, 2004; Bradshaw, 2001; Miller and Cohen, 2001). Various neuroscientists have recognized the overlap between executive functions and concepts drawn from other fields. Convit and his coauthors note, for example, that The characteristics of an individual acting with no forethought and without regard to consequences links the criminologists’ explanation of criminal propensity by inadequate “self-control” (Gottfredson and Hirschi, 1990) or impaired “impulse control” (Wilson and Herrnstein, 1985) and to results suggested by a role of serotonin in impulsive violence (Virkkunen and Linnoila, 1993). The mechanism(s) for impulsive behavior remain unclear. However, most brain researchers would agree that the frontal lobes are crucial in complex tasks when planning is required and that their main function is inhibitory or regulatory (1996:173). Moreover, data from numerous neuroimaging studies vividly show these areas of the brain to be under substantial genetic control.3 Without exception, however, none of the existing criminological studies into the influence of parenting practices on child self-control recognize the possibility that self-control and other executive functions are influenced by genes and by other biological factors, such as environmental tobacco smoke (Yolton, Dietrich, Auinger, Lanphear and Hornung, 2005), blood lead levels (Dietrich, Douglas, Succop, Berger and Bornschein, 2001), or birth complications (Beaver and Wright, 2005).4 Although the evidence 3. We note that the brain is also highly intertwined with its immediate environment. Environmental stimulation aides in synaptogenesis; likewise, a lack of environmental stimulation accelerates neuronal apoptosis. 4. We are not the first to recognize the possibility that low self-control may be under substantial genetic control. Unnever and associates (2003:495) are the exception and note that “…the origins of self-control are not limited to parental 1176 WRIGHT AND BEAVER linking self-regulation to variation in brain structure and functioning is now undeniable, it is still far from clear whether features of the social environment, namely parental socialization practices, influence the development of traits such as low self-control. Far from having a “net effect of zero,” genetic influences may be the dominate influence on executive functions. “The responsible way to tackle the genetic challenge to socialization research,” Caspi and his colleagues argue, “is head on, by using genetically sensitive designs that can provide leverage in identifying environmental risks” (2005:464). Following this advice, we examine the influence of parenting factors on a measure of child self-control in a sample of kindergarten and first-grade students. We also employ, from the same sample, a subsample of twins, and contrast our genetically informed findings against those detected through common SSM assumptions. METHODS SAMPLE Data for this paper come from the Early Childhood Longitudinal Study, Kindergarten Class of 1998–1999 (ECLS-K). Sponsored by the U.S. Department of Education, National Center for Education Statistics (NCES), the ECLS-K is an ongoing study of a nationally representative sample of children designed to assess the impact that the primary schooling years have on learning. The ECLS employed multiple reporting sources to gain detailed information about the children’s behavior, their temperament, their intellectual skills, their social relationships and their environment. Information about such topics was obtained through interviews with the children, their parents and teachers, and school administrators. Four waves of data have been collected thus far: two waves each during kindergarten and first grade. Data collection for wave one began in the fall of 1998, when the children first entered into kindergarten. The second wave of questionnaires were administered the following spring (1999). The last two waves of data were obtained during the fall and spring of the first grade (1999–2000). Only a small subgroup of students, however, were interviewed in the fall of their first grade. Sampling waves assessed during the kindergarten year were measured less than 6 months apart. Some of the parenting measures, moreover, were asked only during the spring term. Given the relatively small time difference between sampling waves, we treat all data collected during the practices…low self-control is not a purely social outcome but is also affected by genetic predispositions.” DO PARENTS MATTER IN SELF-CONTROL? 1177 kindergarten year as from one measurement period. However, we use the measures of child low self-control measured during the spring to maintain temporal ordering. Data from the fall wave of the first grade were excluded from the analysis, making the interval between consecutive measurement periods (kindergarten year to spring term of their first grade), about one year. A unique aspect of the ECLS-K was that when a respondent indicated the presence of a twin, the proband was subsequently included in the sample, netting n=310 twins. Both twins were subject to identical data collection processes and instruments. In terms of cluster size, each twin was reported on by the parent (usually the mother), their teacher and in some instances the mother and the teacher. The total sample size for the ECLS-K includes over 21,000 children. Recall that one purpose of our study was to assess results based on SSM’s to those obtained from a sample of twins. With such a large overall sample, very small differences could easily reach levels of statistical significance. To make our results as valid as possible, we took a random sample of n=1,000 children from the larger sample of 21,000. We chose a sample size of 1,000 for two reasons. First, most national studies have sample sizes that range from 1,000 to 2,000 subjects. Second, no meaningful differences were found in our results when random samples ranging between n=310 and n=2,000 were analyzed. Thus, the results obtained from our random sample of n=1,000 would be the results reported and published without consideration for the twins in the sample. Overall, the ECLS-K is compatible with our research agenda for four main reasons. First, the inclusion of twins permits us to control for the genetic similarity. Second, consistent with Gottfredson and Hirshi’s proposition on the origins of low self-control, a number of parenting questions were asked that tapped into efficacious parenting practices (Wright and Cullen, 2001). Third, the ECLS-K contains childhood measures of low self-control, allowing us to examine the early correlates of low self-control. And, finally, because multiple reporting sources were used, we were able to construct theoretically consistent measures of low self-control based on parent and teacher reports. Taken together, the ECLS-K provides us with a rich data source with which to systematically assess the biological and social origins of low self-control. MEASURES LOW SELF-CONTROL The ECLS-K employed a version of Gresham and Elliott’s (1990) wellknown Social Skills Rating Scale (SSRS), a proprietary assessment 1178 WRIGHT AND BEAVER battery, to measure child self-control. The SSRS is a multi-rater, standardized, normed-referenced assessment battery based on information collected from mothers and teachers. The SSRS contains subscales that tap into child self-control, including overactivity and hyperactivity. The response set for these items were scored 1=never, 2=sometimes, 3=often and 4=very often. Research into the psychometric properties of the SSRS has found the scales and subscales to be high in reliability, moderate in test-retest reliability, and valid (Benes, 1995; Gresham, 2001).5 Because the SSRS is a multi-rater assessment instrument, we created a teacher low self-control scale (wave 2 twin sample alpha=.80; wave 2 random sample alpha=.82; wave 4 twin sample alpha=.82; wave 4 random sample alpha=.85), a parent low self-control scale (wave 2 twin sample alpha=.45; wave 2 random sample alpha=.58; wave 4 twin sample alpha=.57; wave 4 random sample alpha=.59), and a combined low selfcontrol scale (wave 2 twin sample alpha=.61; wave 2 random sample alpha=.63; wave 4 twin sample alpha=.60; wave 4 random sample alpha=.68) for both waves of data. Parental reports, although used widely, are slightly less reliable than information gathered from other sources. Data from teachers, however, have proven to be highly efficient and reliable and help measure conduct that occurs away from parents (Cairns and Cairns, 1994; Harris, 1998). Teachers and parents reported on the child’s ability to manage temper and emotions, on ability to control conduct, and on impulsiveness and activity levels. To test the robustness of our findings, we also employed an expanded measure of low self-control. This scale taps not only into the attention problems outlined, but also into various social problems experienced by children lacking self-control and deficient decision-making processes that often accompany low self-control. This expanded scale contains the following eight items: parental and teacher reports of self-control, parental and teacher reports of approaches to learning, parental reports of the child’s activity level, parental reports of the child’s social interactions with others, teacher reports of the student’s interpersonal skills, and teacher reports of the student’s externalizing problem behaviors (twin sample alpha=.75; random sample alpha=.77). The same measures were used to create the expanded measure of low self-control during first grade (twin 5. The measurement of self-control, or executive control functions in general, is still a matter of substantial debate. This debate has not escaped criminology. In essence, much of the debate centers on whether self-control should be measured attitudinally or through items that capture variation in analogous behaviors. We, however, share the view of Rudolph, Lambert, Clark, and Kurlakowsky. (2001:931) that self-regulation “can be conceptualized as a combination of cognitive, evaluative, and behavioral processes that guide goal-directed action and emotional responsiveness.” DO PARENTS MATTER IN SELF-CONTROL? 1179 sample alpha=.74; random sample alpha=.78). Higher scores on this scale indicate lower levels of self-control. See Appendix A for a description of the variables and scales. Consult Appendix B for descriptive information about the samples. SOCIALIZATION MEASURES Gottfredson and Hirschi maintain that the ways in which parents socialize their children will ultimately decide whether their child develops self-control. In particular, they assert that parents who supervise, recognize and consistently punish childhood transgressions will succeed at instilling self-control in the child. To partially test this perspective, we included five unique measures of parenting behaviors: parental involvement, parental withdrawal, parental affection, physical punishment and family rules. Due to data limitations, we were not able to measure all the parenting variables Gottfredson and Hirschi identify as significant predictors of low self-control. Yet this situation is not unique to the ECLSK data or our analyses in general; rather, prior research has also been hampered by an inability to measure all dimensions of parental socialization (Hay, 2001; Unnever et al., 2003). Even so, many of the measures used in our analysis are consistent with those Gottfredson and Hirschi identify. Parental Involvement. This scale taps into the amount of time the parent spends with the child on various activities. Presumably, while engaging in such activities, parents will also be supervising their child. This nine-item measure was constructed using parental responses to the following questions: how often the parent reads, tells stories, sings songs, helps child with chores, helps with art activities, plays games, teaches the child about nature, builds things and plays sports (twin sample alpha=.74; random sample alpha=.75). Higher scores indicate a greater degree of parental involvement in the child’s life. Parental Withdrawal. The scale serves to capture the degree to which parents retreat from, or hold unfavorable attitudes toward, their child. Nine items comprised this scale: I am too busy to play with child; I have difficulty being warm with the child; being a parent is harder than I anticipated; my child does things that bother me; I have to sacrifice to meet the child’s needs; I feel trapped as a parent; I often feel angry with the child; my child is hard to care for; and being a parent is more work than pleasure (twin sample alpha=.67; random sample alpha=.68). Parental Affection. This four-item scale measures the degree of affection between the child and the parent, and includes the following items: we spend warm, close time together; the child likes me; I always show love for the child; and I express affection to the child. The four items 1180 WRIGHT AND BEAVER were summed, with higher scores representing more parental affection (twin sample alpha=.63; random sample alpha=.59). Physical Punishment. Gottfredson and Hirschi maintain that parents who punish consistently will instill self-control in their child. Although the ECLS-K data do not include measures tapping into the consistency with which parents’ correct misbehavior, two measures do index whether parents would physically punish the child. Parents were presented with a hypothetical scenario, asking them what they would do if the child were to hit them. A list of possible retaliations was then presented, and parents were subsequently asked which, if any, of the punishment strategies they would use. We identified two such actions: if the parent stated she would “hit the child back” or if she would “spank the child.” These two items were then summed, forming the physical punishment index, with higher scores indicating the parent is more likely to resort to physical punishment when faced with disciplining the child. Family Rules. This final socialization measure measures a limited domain of rules within the home. Three questions were asked about rules regarding television viewing. Specifically, parents were asked if there are family rules for which television programs the child can watch, the amount of hours the child is permitted to watch television, and if rules exist on how late or early the child is allowed to watch television. Again, higher scores reflect more family rules (twin sample alpha=.63; random sample alpha=.58). STATISTICAL CONTROLS Gender. Gottfredson and Hirschi maintain that, in general, boys tend to have lower levels of self-control than girls. We therefore include a dichotomous measure of gender in the analyses (1=male; 2=female). Academic Preparedness. Cognitive capacity has been found to be a strong predictor of crime and other behavioral problems (Wilson and Herrnstein, 1985). As such, we included an academic preparedness scale, measured during wave one, to assess the degree to which academic ability is related to low self-control. Children in the ECLS-K were subject to a cognitive assessment battery with three distinct components: language and literacy, mathematical skills, and general knowledge. To ascertain the scores on each of these tests, each child was tested one on one (ECLS-K User’s Manual). To compute the academic preparedness scale, we combined these three scores (twin sample alpha=.87; random sample alpha=.84). The scores tap into the acquisition of knowledge and the child’s preparedness for kindergarten.6 6. Reviewers noted that the measure of academic preparedness overlaps with one of DO PARENTS MATTER IN SELF-CONTROL? 1181 Race. To capture potential differences in self-control between whites and nonwhites, we included race as a control variable. Race was coded (0=white; 1=nonwhite). Neighborhood Disadvantage. Research has found self-control to be influenced by community structural characteristics (Pratt, Turner and Piquero, 2004). To control for the possibility that neighborhood social factors may affect the development of self-control, we included a neighborhood disadvantage scale. Six items, reported by the parents, were summed to form the neighborhood disadvantage scale. Parents were asked how safe it was for their child to play outside, whether garbage and litter were visible on the street, whether there were problems with people selling and using drugs or alcohol in the neighborhood, whether there were problems with burglaries and robberies in the neighborhood, if there were problems with violent crime, and if there were vacant houses nearby. Higher scores indicate more problems in the neighborhood (twin sample alpha=.60; random sample alpha=.73). ANALYTICAL PLAN Harris (1998) offered a stinging critique of standard social science methodologies, noting that most SSMs include data on one child and one parent, inferring from that relationship to other relationships within the home (see also Caspi et al., 2005). She also argues that SSMs cannot account for similarities in genetic commonalities between individuals within the same home. To Harris, most of the current empirical literature on the role and effects of parenting is seriously biased (referenced in Pinker, 2002). We take into account Harris’s criticism by analyzing a random subsample of youth as well as a sample of twins. We also use statistical models unique to each sample. In the random subsample we use OLS regression. This technique does not account for the clustering of Gottfredson and Hirschi’s dimensions of low self-control: preference for physical activities. According to reviewers, we may be committing a tautology primarily because we are predicting low self-control with a component of self-control. The academic preparedness scale does not ask whether students prefer mental activities over physical activities. Rather, the academic preparedness scale tests both what the student has learned, and the student's ability to read and to calculate basic mathematical equations. The measure was originally designed to assess individual differences in school preparedness for children entering Kindergarten. And as Grasmick, Tittle, Bursik and Arneklev (1993) found, the physical preference dimension of low self-control was the weakest correlate with their well-known and widely used scale of self-control. Still, we recalculated all of the models without including the measure of academic preparedness. The results were virtually identical to those reported with the inclusion of the academic preparedness measure. Future research would benefit by examining the nexus between IQ and other measures of intellect, and low self-control. 1182 WRIGHT AND BEAVER individuals within the household; a serious violation of the technique but one that is commonly committed. Genetic similarity within families translates into a loss of statistical independence between observations. OLS regression assumes that observations are independent of each other and thus that the correlation between error terms across observations is zero (0). Of course, children are not randomly assigned within families, and monozygotic twins share all of their genes. Intra-class correlations, (?=rho), a measure of the degree of homogeneity of phenotypes within families, generally range between .20 to .60, indicating that 20 to 60 percent of the variance in the outcome of interest is accounted for by the clustering of observations (Wright and Cullen, 2001). Methodological research, however, has found that ICCs as small as .1 can downwardly bias estimates of the standard errors and in some cases slope estimates (Zyzanski, Flocke and Dickinson, 2004). The result of ignoring genetic similarities within families likely results in the overestimation of significant effects (Type 1 errors), and the inaccurate attribution of “substantive meaning” to parenting variables. In the twin sample we employ a random-effects regression analysis, which is virtually synonymous with hierarchical linear modeling (HLM). This technique allows for the estimation of the proportion of the variance in the dependent variable accounted for by the nonrandom clustering of subjects within twin dyads. In this case, the clustering of subjects occurs due to genetic similarity—that is, monozygotic or dyzygotic twin status. HLM and random-effects models account for the loss of statistical independence and produce robust standard errors through an iterative process. The resulting maximum-likelihood slope estimates are assessed against standard errors that are comparatively more conservative. Unlike studies in which the central concern is in estimating the heritability of a certain trait, our concern is in estimating the potential influences of a range of predictors, controlling in part for genetic similarities. As such, we avoid model fitting techniques that are generally reserved for estimating heritability coefficients.7 Our analysis is similar to Tully, Arseneault, Caspi, Moffitt and Morgan’s analysis of data from the Environmental Risk Longitudinal Twin Study (2004). 7. Model fitting procedures revealed that low self-control is 66 percent heritable. The baseline HLM model produced an ICC of .55, indicating that 55 percent of the variation in self-control could be attributed to the clustering of twins in the same household. The difference between the results from the SEM model and the HLM model indicates that HLM may underestimate the actual amount of clustering in the twin dyads. DO PARENTS MATTER IN SELF-CONTROL? 1183 RESULTS Table 1 shows the results of our OLS and HLM analyses. Looking first at the OLS regression of parental reports of their child’s self-control we found that three of the parenting measures significantly accounted for variation in low self-control: parental withdrawal, parental affection and family rules. As predicted by low self-control theory, parental withdrawal was positively associated with low control, and family rules and parental affection were inversely related to self-control. Gender, academic preparedness and neighborhood disadvantage were also related to low self-control in the theoretically expected direction. These results are precisely what would be predicted from low self-control theory under standard socialization assumptions. Table 1. Effects of Parenting on Child's Low Self-Control in Kindergarten Variables Parental Reports Teacher Reports Total Composite Score OLS HLM OLS HLM OLS HLM Socialization Measures Parental Involvement -.02 (-.73) -.01 (-.86) -.01 (-.31) -.02 (-1.28) -.02 (-.68) -.04 (-1.51) Parental Withdrawal .32* (11.49) .05* (2.60) .08* (2.74) .03 (1.15) .24* (8.48) .08* (2.26) Parental Affection -.09* (-3.32) -.10* (-2.55) -.02 (-.66) .01 (.24) -.07* (-2.50) -.10 (-1.31) Physical Punishment -.01 (-.30) .38* (2.63) .04 (1.24) -.04 (-.21) .02 (.64) .35 (1.33) Family Rules -.08* (-2.97) .09 (1.34) -.05 (-1.68) -.08 (-.89) -.08* (-3.03) .02 (.13) Control Variables Gender -.10* (-3.73) -.16 (-1.55) -.21* (-7.37) -.27* (-2.17) -.20* (-7.48) -.43* (-2.45) Academic Preparedness -.21* (-7.66) -.00 (-1.35) -.16* (-5.46) -.01* (2.29) -.23* (-8.13) -.01* (-2.33) Race .00 (.10) -.02 (-.11) .03 (.89) .19 (1.07) .02 (.79) .16 (.63) Neighborhood Disadvantage .07* (2.46) .07 (1.25) -.03 (-.94) -.00 (-.06) .02 (.66) .06 (.59) Number of Significant Parenting Parameters 3 3 1 0 3 1 Intra-cluster Correlation .284 .457 .367 * Parameter estimate at least twice its standard error The HLM/twin model, however, reveals a slightly different pattern. Parental withdrawal and parental affection retained statistical significance. 1184 WRIGHT AND BEAVER However, whereas the measure of family rules dropped out of statistical significance, the measure of physical punishment reached it. Moreover, none of the control measures were significant predictors of self-control, nor was neighborhood disadvantage. Evidence from the first set of equations indicates that the OLS/SSM model likely overestimated the number of significant independent effects on parental measures of child self-control. In the HLM model, with the clustering due to genetic similarity accounted for, only three predictors reached significance compared to six in the OLS/SSM model. Teacher reports of child low self-control reveal a very different pattern of results. In the OLS model, the only significant predictors were parental withdrawal, gender and academic preparedness. However, in the HLM model, none of the parenting measures were significantly associated with low self-control. Consistent with the OLS model, gender and academic preparedness were also significant predictors. Finally, examining the composite measure of low self-control we find that, once again, the OLS model overestimated the number of significant parenting effects. In the OLS model, parental withdrawal, parental affection and family rules were significant predictors. In the HLM/twin model, however, only parental withdrawal was. In both models, gender and academic preparedness were significant. The results from Table 1 indicate that traditional social research methodologies tend to inflate the effects parents have on their offspring’s self-control, especially if the reporting source is a parent. We note that because many of the parenting measures were taken contemporaneously with the measures of low self-control, that these models offer the highest likelihood of detecting parenting effects. Still, once clustering due to genetic similarity was controlled, most of the parenting effects dissipated to statistical insignificance. Table 2 presents the results for child self-control, measured one year later when the children were in the first grade, regressed on the various parenting measures. Looking first at the OLS regression of parental reports we find that parental withdrawal, parental affection and family rules were significant predictors of child self-control. So too were gender, academic preparedness and neighborhood disadvantage. Even so, in the HLM/twin model only parental withdrawal and academic preparedness were significantly associated with child low self-control. Once again, the OLS/SSM model overestimated the number of significant parameters. Six parameters reached significance in the OLS model, compared to just two in the HLM/twin model. An even greater contrast can be seen when teacher reports are examined. Only one of the independent variables was associated with low self-control in the HLM/twin analyses—parental withdrawal. Yet in the DO PARENTS MATTER IN SELF-CONTROL? 1185 OLS model, parental withdrawal, family rules, gender and academic preparedness each made independent contributions. In the final set of equations in Table 2 we find that in the OLS model, parental withdrawal, family rules, gender and academic preparedness accounted for variation in low self-control. In the HLM/twin model, parental withdrawal, academic preparedness and race were significantly associated with low self-control. Table 2. Effects of Parenting on Child's Low Self-Control in First Grade Variable Parental Reports Teacher Reports Total Composite Score OLS HLM OLS HLM OLS HLM Socialization Measures Parental Involvement -.02 (-.75) -.01 (-.58) -.04 (-1.17) -.01 (-.63) -.04 (-1.13) -.02 (-.69) Parental Withdrawal .25* (8.49) .07* (3.41) .09* (2.81) .02 (.65) .20* (6.33) .06 (1.80) Parental Affection -.07* (-2.34) -.08 (-1.92) .04 (1.24) .05 (.88) -.02 (-.73) -.05 (-.76) Physical Punishment .01 (.07) .07 (.44) .01 (.33) .33 (1.80) .02 (.72) .38 (1.52) Family Rules -.09* (-3.31) .00 (.01) -.07* (-2.30) -.22* (-2.27) -.10* (-3.30) -.28* (-2.07) Control Variables Gender -.07* (-2.37) -.15 (-1.53) -.21* (-6.94) -.11 (-.87) -.18* (-6.07) -.29 (-1.65) Academic Preparedness -.20* (-6.80) -.01* (-3.34) -.20* (-6.48) -.01 (-1.88) -.26* (-8.45) -.01* (-3.55) Race .04 (1.29) .24 (1.69) .02 (.47) .29 (1.68) .03 (1.10) .46* (1.93) Neighborhood Disadvantage .06* (2.19) -.02 (-.29) -.01 (-.29) .00 (.00) .02 (.75) -.02 (-.16) Number of Significant Parenting Parameters 3 1 2 1 2 1 Intra-cluster Correlation .402 .385 .315 * Parameter estimate at least twice its standard error Consistent with the findings presented in Table 1, the results shown in Table 2 indicate that standard OLS/SSM models overestimate the influence of various variables on self-control. Of particular interest however, is the overestimation of parenting influences. Our parenting measures were consistently related to child self-control in the OLS/SSM models in general, but were more consistently related when parents were used as the reporting source. This bias, we note, is the type predicted by Harris (1998) and Pinker (2002). Once similarities due to genetic 1186 WRIGHT AND BEAVER influences were removed, the effects of other variables, particularly parenting variables, were reduced or eliminated. In our last set of analyses, depicted in Table 3, we use an expanded measure of low self-control. Consistent with our prior analyses, we examine a contemporaneous measure of low self-control and a prospective measure of low self-control. Table 3. Effects of Parenting on Full Measures of Child's Low Self-Control Variable (Kindergarten) (First Grade) OLS HLM OLS HLM Socialization Measures Parental Involvement -.06* (-2.29) -.14* (-3.23) -.10* (-3.18) -.05 (-1.05) Parental Withdrawal .21* (7.59) .06 (.91) .16* (5.20) .10 (1.42) Parental Affection -.10* (-3.42) -.20 (-1.44) -.03 (-1.06) -.07 (-.50) Physical Punishment .03 (1.01) .25 (.51) -.01 (-.04) .22 (.45) Family Rules -.05 (-1.77) -.10 (-.40) -.10* (-3.25) -.46 (-1.77) Control Variables Gender -.23* (-8.84) -.93* (-2.77) -.22* (-7.59) -.86* (-2.54) Academic Preparedness -.34* (-12.42) -.04* (-5.51) -.37* (-12.43) -.04* (-5.20) Race .01 (.26) .17 (.38) .01 (.14) .70 (1.52) Neighborhood Disadvantage -.01 (-.30) .10 (.54) .01 (.03) .05 (.23) Number of Significant Parenting Parameters 3 1 3 0 Intra-cluster Correlation .328 .374 * Parameter estimate at least twice its standard error Parental involvement, parental withdrawal, parental affection, gender and academic preparedness made independent contributions to the explained variance in child low self-control in the OLS model (kindergarten). However, consistent with prior analyses, only parental withdrawal, gender and academic preparedness were significantly related to low self-control in the HLM/twin model. A similar pattern was again detected when we regressed the prospective measure of child self-control on the independent predictors. In the OLS model, parental involvement, parental withdrawal, family rules, gender and academic preparedness were significant predictors. In the HLM/twin model, though, none of the parenting measures were significantly related DO PARENTS MATTER IN SELF-CONTROL? 1187 to child self-control. Only gender and academic preparedness retained statistical significance. DISCUSSION Our study sought to answer a single question: do parents matter in the etiology of self-control? At first glance, this would appear to be a rather simple empirical question. Empirical research, however, has a bad habit of exposing the underlying complexities of even the simplest research question. So it was in this instance. On close inspection, we realized that an answer to this question would require data that allowed for the estimation of genetic similarity between twins and their mother, data that contained measures of various parenting constructs and data that contained multiple reporting sources. The ECLS-K fit these requirements. We address the findings from this study along two intertwined dimensions: substantive findings and methodological implications. In general, our analyses revealed that parenting variables were inconsistently and weakly related to contemporaneous measures of child self-control in kindergarten, and were inconsistently related to prospective measures of self-control in the first grade. Parenting influences typically reached statistical significance in the OLS models, models that did not account for the clustering of responses due to genetic similarity. We also note that parenting influences reached significance only when parental reports were used. When teacher reports of child self-control were analyzed, parenting features were consistently unrelated to child self-control. Teachers, we note, observe children under very difference circumstances than parents; circumstances that usually require young children to exhibit self-control (Cairns and Cairns, 1994). Whether parenting matters in the etiology of child self-control thus appears to be deeply intertwined with the type of methodology and analyses employed. Standard social methodologies (SSMs), which usually measure only one child and a parent within a household and which typically ignore genetic similarities between subjects, appear to overestimate the influence of parenting on child self-control. The results of our OLS models, derived from standard SSM assumptions and methodologies, would generally be accepted as evidence linking parenting to self-control in children. Employing a genetically informed methodology and analysis, however, alters, or at least conditions, such a conclusion. Overall, employing a more rigorous methodology and analysis reduced or eliminated the influence of several variables, most noticeably the parenting variables. Even here, however, it appears that methodology makes a difference. Parenting effects were detected only when parent reports of child self-control were 1188 WRIGHT AND BEAVER used. The use of teacher reports in our twin/HLM models detected no significant association between parenting variables and child self-control. This result held contemporaneously and longitudinally. The use of parent reports, even in our twin sample, confounds the association between parenting behaviors and the child’s self-control. That is, parents who are far removed from the daily happenings of their children may be likely to report self-control problems in their children. This effect may be “real,” but it may also reflect individual differences shared between parents and their offspring, or at least differences rooted in behavioral and attitudinal factors that vary between parents. Either way, the use of parent reports may overestimate the influence that parenting behaviors have on a child’s traits. Evidence from the teacher report models strongly suggests this to be the case. HOW SHOULD OUR FINDINGS BE INTERPRETED? The contrast in findings begs the question of what the appropriate benchmark to evaluate the validity of our findings is. Some readers will point to the OLS models as offering evidence sufficient to demonstrate parenting effects. They may also point out that the valid conclusions about parenting behaviors can be derived only from cross-sectional analyses. However, more conservative analysts will point out that the OLS estimates are substantially biased and do not meet the criteria outlined by Harris to reject the null hypothesis that parents “don’t matter.” Nor do the parenting variables generate consistent effects when temporal order is specified correctly. What, then, do we make of our findings? We suspect the most valid and conservative findings are those that emerge from the composite measure of self-control in the twin/HLM sample. These are consistent across time and indicate that the measure of parental withdrawal, which captures variation in the extent to which parents report feelings of stress and emotional distance from their children, is associated with increases in child low self-control contemporaneously and over a one-year period. The effects are only marginally significant, however, which indicates that the true effect size may be trivial. To buttress our initial findings, we also merged the random sample and the twin sample. The results are shown in Appendix C and converge with those generated by analyzing the samples separately. As such, these findings provide tangible evidence in favor of Harris’s (1998) proposition that, net of genetic similarities within households, parental socialization techniques minimally influence the individual traits of their children.8 8. We also note that Gottfredson and Hirschi claim that self-control materializes prior to the age of twelve and becomes very unlikely to appear thereafter. Neurological studies, however, dispute this possibility by showing that cerebral volume increases DO PARENTS MATTER IN SELF-CONTROL? 1189 Does this mean that parents do not matter? Of course not, nor do we encourage readers to blithely accept that possibility given our analyses. Instead, our view is that the influence of parental socialization factors, as well as other environmental features, are conditioned by the genotypes of the parent and child. Recent evidence by Caspi and his colleagues, for example, has documented how individual responses to life-stresses and to experiencing parental abuse vary by child genotype (Caspi et al., 2002; Caspi et al., 2003). As behavioral geneticists frequently tell us, children within the same family are often very different. These differences cannot be explained by factors that do not vary between children, differences that aggregated parenting measures are not capable of assessing (Plomin, 1995). Parents likely influence their children in ways that are more complicated than is typically assumed. Parents may moderate the influence of specific child traits (Tully et al., 2004), or the traits of parents may interact in unique ways with the traits of each of their children. Parents may also create environments that are so bleak and abusive that the environmental effects overshadow any genetic influences (Harris, 1998). Moreover, the traits of the child likely influence the reactions of parents.9 Difficult children constantly challenge parental authority and limit-setting efforts. In either case, whatever influence parents have on the traits of their children, it likely will involve a more sophisticated understanding of the genetically influenced, mutually dynamic relationships that occur within households. Although other studies will have to verify our findings, we note that recent neuroimaging research and studies of twins provide strong evidence of the genetic influence on brain structure and functioning, especially as it relates to executive control functions (Barkley, 1997; Thompson et al., 2001). Through the use of complex imaging, Thompson and his colleagues (2001) analyzed brain structural differences in a sample of monozygotic and dizygotic twins. Their results graphically depict the strong genetic foundation underpinning brain growth and functioning. Particularly striking was their analysis of monozygotic twins. This analysis revealed almost identical gray matter volume in the frontal lobes of the twins and in through the age of 20, while temporal grey matter increases in volume through the age of 16.5 in males and 16.7 in females (Giedd, Blumenthal, Jeffries, Castellanos, Liu, Zijdenbos, Paus, Evans, and Rapoport, 1999). Future research should examine variation in self-control across a longer age-range, as well as include a broader array of socialization measures. 9. As one reviewer noted, the effects generated from the measure of parental withdrawal could represent the influence of the child on the parent. In subsequent analyses, we included a measure of prior low self-control (results not shown). Inclusion of that measure reduced the effect of parental withdrawal to statistical insignificance. In the language of behavioral genetics, this dynamic is known as a provocative gene X environment interaction. 1190 WRIGHT AND BEAVER regions that control language acquisition. Other imaging studies have found that genetic heritability in brain structure is evident through the seventh and eighth decades of life (Pfefferbaum, Sullivan, Swan, and Carmelli, 2000). SELF-CONTROL AND CRIMINOLOGY Neuroimaging studies highlight the close correspondence between specific traits, such as low self-control, and the genetic controls that direct neuronal growth in specific brain structures related to specific traits. More important, they help to explain why common environmental features, such as parental socialization efforts, have only modest to trivial effects (Wright and Cullen, 2001). At least as it relates to self-control, once similarities in genetic heritability are accounted for, the range of variance left to explain is quite restricted. As it applies to antisocial behavior in young children, Arseneault and her colleagues (2003) found that in a sample of 1116 twin pairs, genetic factors accounted for 82 percent of the variance, while experiential factors accounted for only 18 percent of the variance. Recent meta-analytic reviews also support these findings (Miles and Carey, 1997). Gottfredson and Hirschi openly exclude the possibility that self-control has a genetic base. Our study, along with others from various fields, suggests that for self-control theory to be a valid theory of crime it must incorporate a more sophisticated understanding of the origins of selfcontrol (Pratt, Turner and Piquero, 2004; see also Pratt et al., 2002). With the caveats that we have not measured all parental socialization practices and that our data are restricted in age range, the pattern of findings reported here contradicts Gottfredson and Hirschi’s hypotheses that link parental socialization to the development of self-control. Excluding genetic influences on self-control and related traits likely misspecifies a theory that, in general, has gained widespread empirical support (Pratt and Cullen, 2000). It also fails to recognize the large body of research that has linked executive control functions to a range of biological factors, such as prenatal maternal cigarette smoking and drug use (Gibson and Tibbetts, 1998), prenatal lead absorption (Bellinger, Leviton, Allred and Rabinowitz, 1994), and in utero anoxia (Beaver and Wright, 2005). We suggest that self-control theory be revised to incorporate this body of literature. Lastly, we note that methodological assumptions shape research findings, and in this case, it is clear to see that the findings from the OLS models overestimated the influence of a range of factors, including parenting and neighborhood variables. Although we do not believe that our methodology is the only way to analyze socialization hypotheses, we do suggest that it is time for criminologists to come to grips with the DO PARENTS MATTER IN SELF-CONTROL? 1191 confluence of genetic similarities and individually specific experiences (Caspi et al., 2005). There is now a voluminous literature that links biological features to human growth and social development. To the extent that insights from this body of knowledge remain excluded, criminology will remain on the periphery of other, more established disciplines (Daly and Wilson, 1988; Rowe, 1994; Udry, 1995; Walsh and Ellis, 2004; Wilson, 1998). REFERENCES Aron, Adam R., Trevor W. Robbins and Russell A. Poldrack 2004 Inhibition and the right inferior frontal cortex. Trends in Cognitive Sciences 8: 170–177. Arseneault, Louise, Terrie E. Moffitt, Avshalom Caspi, Alan Taylor, Fruhling V. Rijsdijk, Sara R. Jaffee, Jennifer C. Ablow and Jeffrey R. Measelle 2003 Strong genetic effects on cross-situational antisocial behaviour among 5-year-old children according to mothers, teachers, examiner-observers, and twins’ self-reports. Journal of Child Psychology and Psychiatry 44: 832–848. Barkley, Russell A. 1997 ADHD and the Nature of Self-Control. New York: Guilford Press. 2005 Dr. Russell Barkley on AD/HD: Excerpts from his lecture in San Francisco, CA on June 17, 2000. Available online at: http://www.schwablearning.org/pdfs/2200_7- barktran.pdf?date=4-12-05 Beaver, Kevin M. and John Paul Wright 2005 Evaluating the effects of birth complications on low self-control in a sample of twins. International Journal of Offender Therapy and Comparative Criminology 49: 450–471. Bellinger, David, Alan Leviton, Elizabeth Allred and Michael Rabinowitz 1994 Pre- and postnatal lead exposure and behavior problems in school-aged children. Environmental Research 66: 12–30. Benes, Kathryn M. 1995 Review of the social skills rating system. In Conoley and Impara (eds.), Twelfth Mental Measurements Yearbook. Lincoln, NE: Buros Institute of Mental Measurement. Bradshaw, John L. 2001 Developmental Disorders of the Frontostriatal System: Neuropsychological, Neuropsychiatric and Evolutionary Perspectives. Philadelphia, PA: Taylor and Francis. 1192 WRIGHT AND BEAVER Cairns, Robert D. and Beverley D. Cairns 1994 Lifelines and Risks: Pathways of Youth in Our Time. Cambridge: Cambridge University Press. Caspi, Avshalom, Brent W. Roberts and Rebecca L. Shiner 2005 Personality development: Stability and change. Annual Review of Psychology 56: 453–484. Caspi, Avshalom, Joseph McClay, Terrie E. Moffitt, Jonathan Mill, Judy Martin, Ian W. Craig, Alan Taylor and Richie Poulton 2002 Role of genotype in the cycle of violence in maltreated children. Science 297: 851–854. Caspi, Avshalom, Karen Sugden, Terrie E. Moffitt, Alan Taylor, Ian W. Craig, HonaLee Harrington, Joseph McClay, Jonathan Mill, Judy Martin, Antony Braithwaite and Richie Poulton 2003 Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene. Science 301: 386–389. Cochran, John K., Peter B. Wood, Christine S. Sellers, Wendy Wilkerson and Mitchell B. Chamlin 1998 Academic dishonesty and low self-control: An empirical test of A General Theory of Crime. Deviant Behavior 19: 227–255. Cohen, David B. 1999 Strangers in the Nest: Do Parents Really Shape Their Child’s Personality, Intelligence, or Character? New York: John Wiley & Sons. Convit, Antonio, Richard Douyon, Kathy F. Yates, Gwenn Smith, Pal Czobor, Jose de Asis, Joseph Vitrai, Louis Camus and Jan Volavka 1996 Frontotemporal abnormalities and violent behavior. In D. M. Stoff and R. B. Cairns (eds.), Aggression and Violence: Genetic, Neurobiological, and Biosocial Perspectives. Mahwah, NJ: Lawrence Erlbaum. Daly, Martin and Margo Wilson 1988 Homicide. New York: Aldine de Gruyter. Dietrich, Kim N., M. Douglas Ris, Paul A. Succop, Omer G. Berger and Robert L. Bornschein 2001 Early exposure to lead and juvenile delinquency. Neurotoxicology and Teratology 23: 511–518. Dunn, Judy and Robert Plomin 1990 Separate Lives: Why Siblings are so Different. New York: Basic Books. DO PARENTS MATTER IN SELF-CONTROL? 1193 Feldman, S. Shirley and Daniel A. Weinberger 1994 Self-restraint as a mediator of family influences on boys’ delinquent behavior: A longitudinal study. Child Development 65: 195–211. Geis, Gilbert 2000 On the absence of self-control as the basis for a general theory of crime: A critique. Theoretical Criminology 4: 35–53. Gibbs, John J., Dennis Giever and George E. Higgins 2003 A test of Gottfredson and Hirschi’s generally theory using structural equation modeling. Criminal Justice and Behavior 30: 441–458. Gibbs, John J., Dennis Giever and Jamie S. Martin 1998 Parental management and self-control: An empirical test of Gottfredson and Hirschi’s general theory. Journal of Research in Crime and Delinquency 35: 40–70. Gibson, Chris L. and Stephen G. Tibbetts 1998 Interaction between maternal cigarette smoking and apgar scores in predicting offending behavior. Psychological Reports 83: 579–586. Giedd, Jay N., Jonathan Blumenthal, Neal O. Jeffries, F. X. Castellanos, Hong Liu, Alex Zijdenbos, Tomas Paus, Alan C. Evans and Judith L. Rapoport 1999 Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neuroscience 2: 861–863. Gottfredson, Michael and Travis Hirschi 1990 A General Theory of Crime. Palo Alto, CA: Stanford University Press. Grasmick, Harold G., Charles R. Tittle, Robert J. Bursik, Jr. and Bruce J. Arneklev 1993 Testing the core empirical implications of Gottfredson and Hirschi’s general theory of crime. Journal of Research in Crime and Delinquency 30: 5–29. Gresham, Frank M. 2001 Assessment of social skills in children and adolescents. In Jac J.W. Andrews, Donald H. Saklofske, and Henry L. Janzen (Eds.), Handbook of Psychoeducational Assessment:Ability, Achievement, and Behavior in Children. New York: Academic Press. 1194 WRIGHT AND BEAVER Gresham, Frank M. and Stephen N. Elliott 1990 The Social Skills Rating System. Circle Pines, MN: American Guidance Service. Harris, Judith R. 1998 The Nurture Assumption: Why Children Turn Out the Way They Do. New York: The Free Press. Hay, Carter 2001 Parenting, self-control, and delinquency: A test of self-control theory. Criminology 39: 707–736. Higgins, George E. 2002 General theory of crime and deviance: A structural equation modeling approach. Journal of Crime and Justice 25: 71–95. Loeber, Rolf and Magda Stouthamer-Loeber 1986 Family factors as correlates and predictors of juvenile conduct problems and delinquency. In Michael Tonry and Norval Morris (eds.), Crime and Justice: An Annual Review of Research. Vol. 7. Chicago: University of Chicago Press. Lynskey, Dana Peterson, L. Thomas Winfree, Finn-Aage Esbensen and Dennis L. Clason 2000 Linking gender, minority group status, and family matter to selfcontrol theory: A multivariate analysis of key self-control concepts in a youth-gang context. Juvenile and Family Court Journal 51: 1–19. Marcus, Bernd 2004 Self-control in the general theory of crime: Theoretical implications of a measurement problem. Theoretical Criminology 8: 33–55. Mick, Eric, Joseph Biederman, Jefferson Prince, Marianna J. Fischer and Stephen V. Faraone 2002 Impact of low birth weight on attention-deficit hyperactivity disorder. Developmental and Behavioral Pediatrics 23: 16–22. Miles, Donna R. and Gregory Carey 1997 Genetic and environmental architecture of human aggression. Journal of Personality and Social Psychology 72: 207–217. Miller, Earl K. and Jonathan D. Cohen 2001 An integrative theory of prefrontal cortex function. Annual Review of Neuroscience 24: 167–202. DO PARENTS MATTER IN SELF-CONTROL? 1195 National Institute of Mental Health 2003 Attention Deficit Hyperactivity Disorder. Bethesda, MD: National Institute of Mental Health, National Institutes of Health, US Department of Health and Human Services. Available from: http://www.nimh.nih.gov/publicat/adhd.cfm Neiderhiser, Jenae, M., David Reiss, E. Mavis Hetherington and Robert Plomin 1999 Relationships between parenting and adolescent adjustment over time: Genetic and environmental contributions. Developmental Psychology 35: 680–692. O’Connor, Thomas G., Jenae M. Neiderhiser, David Reiss, E. Mavis Hetherington and Robert Plomin 1998 Genetic contributions to continuity, change, and the cooccurrence of antisocial and depressive symptoms in adolescence. Journal of Child Psychology and Psychiatry and Allied Disciplines 39: 323–336. Patterson, Gerald R. 1982 Coercive Family Process. Eugene, OR: Castilia. Pinker, Steven 2002 The Blank Slate: The Modern Denial of Human Nature. New York: Viking. Pfefferbaum, Adolf, Edith V. Sullivan, Gary E. Swan and Dorit Carmelli 2000 Brain structure in men remains highly heritable in the seventh and eighth decades of life. Neurobiology of Aging 21: 63–74. Plomin, Robert 1995 Genetics and children’s experiences in the family. Journal of Child Psychology and Psychiatry 36: 33–68. Plomin, Robert, Michael J. Owen and Peter McGuffin 1994 The genetic basis of complex human behaviors. Science 264:1733-1739. Polakowski, Michael 1994 Linking self- and social control with deviance: Illuminating the structure underlying a general theory of crime and its relation to deviant identity. Journal of Quantitative Criminology 10: 41–78. Pratt, Travis C. and Francis T. Cullen 2000 The empirical status of Gottfredson and Hirschi’s general theory of crime: A meta-analysis. Criminology 38: 931–964. 1196 WRIGHT AND BEAVER Pratt, Travis C., Michael G. Turner and Alex R. Piquero 2004 Parental socialization and community context: A longitudinal analysis of the structural sources of low self-control. Journal of Research in Crime and Delinquency 41: 219–243. Pratt, Travis C., Francis T. Cullen, Kristie R. Blevins, Leah Daigle and James D. Unnever 2002 The relationship of attention deficit hyperactivity disorder to crime and delinquency: A meta-analysis. International Journal of Police Science and Management 4: 344–360. Price, Thomas S., Emily Simonoff, Irwin Waldman, Philip Asherson and Robert Plomin 2001 Hyperactivity in preschool children is highly heritable. Journal of the American Academy of Child and Adolescent Psychiatry 12: 1362–1364. Reiss, David, Nenae M. Neiderhiser, E. Mavis Hetherington and Robert Plomin 2000 The Relationship Code: Deciphering the Genetic and Social Influences on Adolescent Development. Cambridge: Cambridge University Press. Rietveld, M.J.H., J. J. Hudziak, M. Bartels, C.E.M. van Beijsterveldt and D. I. Boomsma 2003 Heritability of attention problems in children: Cross-sectional results from a study of twins, age 3 to 12 years. Neuropsychiatric Genetics 1176: 102–113. Rowe, David C. 1994 The Limits of Family Influence: Genes, Experience, and Behavior. New York: Guilford. Rudolph, Karen D., Sharon F. Lambert, Alyssa G. Clark and Kathryn D. Kurlakowsky 2001 Negotiating the transition to middle school: The role of selfregulatory processes. Child Development 72: 929–946. Sampson, Robert J. and John H. Laub 1995 Understanding variability in lives through time: Contributions of life-course criminology. Studies on Crime and Crime Prevention 4: 143–158. Scarr, Sandra and Kathleen McCartney 1983 How people make their own environments: A theory of genotype?environment effects. Child Development 54: 424–435. DO PARENTS MATTER IN SELF-CONTROL? 1197 Spencer, Thomas J., Joseph Biederman, Timothy E. Wilens and Stephen V. Faraone 2002 Overview and neurobiology of attention-deficit/hyperactivity disorder. Journal of Clinical Psychiatry 63:3-9. Thompson, Paul, Tyrone D. Cannon, Katherine L. Narr, et al. 2001 Genetic influences on brain structure. Neuroscience 4: 1–6. Tully, Lucy A., Louise Arseneault, Avshalom Caspi, Terrie E. Moffitt and Julia Morgan 2004 Does maternal warmth moderate the effects of birth weight on twins’ attention-deficit/hyperactivity disorder (ADHD) symptoms and low IQ? Journal ofConsulting and Clinical Psychology 72: 218–226. Udry, Richard J. 1995 Sociology and biology: What biology do sociologists need to know? Social Forces 73: 1267–1278. Unnever, James D., Francis T. Cullen and Travis C. Pratt 2003 Parental management, ADHD, and delinquent involvement: Reassessing Gottfredson and Hirschi’s general theory. Justice Quarterly 20: 471–500. Virkkunen Matti and Markku Linnoila 1993 Serotonin in personality disorders with habitual violence and impulsivity. In S. Hodgins (ed.), Mental Disorder and Crime. Newbury Park, CA: Sage Publications. Walsh, Anthony 2002 Biosocial Criminology: Introduction and Integration. Cincinnati, OH: Anderson. Walsh, Anthony and Lee Ellis 2004 Ideology: Criminology’s Achilles’ heel? Quarterly Journal of Ideology 27: 1–25. Wilson, Edward O. 1998 Consilience: The Unity of Knowledge. New York: Alfred A. Knopf. Wilson, James Q. and Richard Herrnstein 1985 Crime and Human Nature: The Definitive Study of the Causes of Crime. New York: Simon and Schuster. Wright, John Paul and Francis T. Cullen 2001 Parental efficacy and delinquent behavior: Do control and support matter? Criminology 39: 677–706. 1198 WRIGHT AND BEAVER Yolton, Kimberly, Kim Dietrich, Peggy Auinger, Bruce P. Lanphear and Richard Hornung 2005 Exposure to environmental tobacco smoke and cognitive abilities among U.S. children and adolescents. Environmental Health Perspectives 113: 98–103. Zyzanski, Stephen G., Susan A. Flocke and L. Miriam Dickinson 2004 On the nature and analysis of clustered data. Journal of Family Medicine 2: 199–200. John Paul Wright is associate professor of criminal justice at the University of Cincinnati. He has published articles on the effects of adolescent employment on delinquency, the effects of parenting on offspring misbehavior, and on the role of money in youthful misconduct. His current research focuses on the genetic heritability of traits related to crime, including self-control, stability in antisocial behavior over time, and the biosocial development of serious violence. Kevin M. Beaver is a doctoral candidate in the Division of Criminal Justice at the University of Cincinnati. He received his BA in sociology from Ohio University and his MS in criminal justice from the University of Cincinnati. His research and teaching interests include life-course and biosocial criminology, the heritability of antisocial behavior, and the stability of violent offending. He has published in the Journal of Quantitative Criminology, Youth Violence and Juvenile Justice, and International Journal of Offender Therapy and Comparative Criminology. DO PARENTS MATTER IN SELF-CONTROL? 1199 APPENDIX A. DESCRIPTION OF VARIABLES AND SCALES Low Self-Control Scales Teacher Ratings of Low Self-Control in Kindergarten (wave 2) Scale created by summing the following items: Teacher reports of the student’s ... 1) externalizing problem behaviors 2) self-control Teacher Ratings of Low Self-Control in First Grade (wave 4) Scale created by summing the following items: Teacher reports of the student’s ... 1) externalizing problem behaviors 2) self-control Parental Ratings of Low Self-Control in Kindergarten (wave 2) Scale created by summing the following items: Parental reports of the child’s ... 1) self-control 2) impulsivity Parental Ratings of Low Self-Control in First Grade (wave 4) Scale created by summing the following items: Parental reports of the child’s ... 1) self-control 2) impulsivity Total Composite Score of Low Self-Control in Kindergarten (wave 2) Scale created by summing the following items: Teacher reports of the student’s ... 1) externalizing problem behaviors 2) self-control 3) approaches to learning (e.g., attentiveness and persistence) 4) interpersonal skills (e.g., ability to form friendships) Parental reports of the child’s ... 5) self-control 6) impulsivity 7) approaches to learning (e.g., concentration and persistence) 8) social interactions (e.g., positive interactions with peers) Total Composite Score of Low Self-Control in First Grade (wave 4) Scale created by summing the following items: Teacher reports of the student’s ... 1) externalizing problem behaviors 2) self-control 3) approaches to learning (e.g., attentiveness and persistence) 4) interpersonal skills (e.g., ability to form friendships) 1200 WRIGHT AND BEAVER Parental reports of the child’s ... 5) self-control 6) impulsivity 7) approaches to learning (e.g., concentration and persistence) 8) social interactions (e.g., positive interactions with peers) Socialization Measures Parental Involvement in Kindergarten (wave 1) Scale created by summing the following items (according to parental reports): How often the parent ... 1) reads to the child 2) tells the child stories 3) sings songs with the child 4) helps the child with art activities 5) helps the child with chores 6) plays games with the child 7) teaches the child about nature 8) helps the child build things 9) plays sports with the child Parental Withdrawal in Kindergarten (wave 2) Scale created by summing the following items (according to parental reports): 1) Does the parent have to sacrifice to meet the child’s needs 2) Does the respondent feel trapped as a parent 3) Is the parent too busy to spend time with the child 4) Does the parent often feel angry with the child 5) Is it hard for the parent to be warm to the child 6) Is the child harder to care for than anticipated 7) Being a parent is harder than the parent expected 8) Does the child do things that bother the parent 9) Is being a parent more work than pleasure Parental Affection in Kindergarten (wave 2) Scale created by summing the following items (according to parental reports): 1) Does the parent and child spend warm, close time together 2) Does the child like the parent 3) Does the parent always show love for the child 4) Does the parent express affection to the child Physical Punishment in Kindergarten (wave 2) Index created by summing the following items (according to parental reports): If the child hit the parent, the parent would ... 1) hit the child back 2) spank the child DO PARENTS MATTER IN SELF-CONTROL? 1201 Family Rules (wave 2) Scale created by summing the following items (according to parental reports): Are there family rules ... 1) for which television programs the child can watch 2) limiting the number of hours the child can watch television 3) pertaining to how early/late the child can watch television Scale Included for Statistical Control Neighborhood Disadvantage Scale (wave 2) Scale created by summing the following items (according to parental reports): 1) How safe is it for the child to play outside 2) Is there garbage and litter on the street 3) Are there problems with people using or selling drugs in the neighborhood 4) Are there problems with burglaries or robberies in the neighborhood 5) Are there problems with violent crime in the neighborhood 6) Are there vacant houses in the neighborhood Appendix B. ECLS-K Descriptive Statistics Twin Sample Random Sample Combined Sample Mean SD Mean SD Mean SD Socialization Measures Parental Involvement 24.72 4.60 25.10 4.51 25.03 4.52 Parental Withdrawal 11.73 3.04 12.15 3.25 12.08 3.22 Parental Affection 14.66 1.59 14.76 1.53 14.74 1.54 Physical Punishment .20 .40 .22 .43 .22 .43 Family Rules 2.23 .91 2.23 .90 2.24 .90 Control Variables Percentage Male 40% 51% 49% Academic Preparedness 147.16 26.11 152.44 25.37 151.56 25.56 Percentage White 63% 56% 57% Neighborhood Disadvantage 17.34 3.04 17.10 1.56 17.15 1.50 Kindergarten Parental Reports 3.84 .88 4.05 1.01 4.02 1.00 Teacher Reports 3.25 1.05 3.48 1.14 3.44 1.13 Composite Score 7.08 1.51 7.51 1.69 7.43 1.66 Full Measure 15.68 2.92 16.18 3.13 16.09 3.10 First-Grade Parental Reports 3.93 .88 3.92 .99 3.92 .97 Teacher Reports 3.24 1.01 3.49 1.16 3.45 1.14 Composite Score 7.0 1.49 7.39 1.71 7.34 1.68 Full Measure 15.75 2.80 16.21 3.17 16.13 3.12 1202 WRIGHT AND BEAVER Appendix C. Effects of Parenting on Full Measures in Combined Sample Variable (Kindergarten) (First Grade) OLS HLM OLS HLM Socialization Measures Parental Involvement -.06* (-3.51) -.14* (-3.52) -.07* (-3.49) -.04 (-1.05) Parental Withdrawal .19* (7.42) .05 (.91) .16* (5.62) .12* (2.03) Parental Affection -.19* (-3.80) -.25* (-2.09) -.07 (-1.27) -.13 (-1.08) Physical Punishment .21 (1.18) .35 (.81) -.00 (-.00) .32 (.75) Family Rules -.19* (-2.24) -.25 (-1.13) -.40* (-4.15) -.45 (-1.88) Control Variables Gender -1.36* (-9.42) -.94* (-2.83) -1.36* (-8.44) -1.05* (-3.09) Academic Preparedness -.04* (-13.67) -.04* (-6.08) -.05* (-13.21) -.04* (-5.28) Race .09 (.58) .14 (.34) .18 (1.02) .66 (1.61) Neighborhood Disadvantage -.01 (-.12) -.09 (-.60) -.02 (-.33) -.03 (-.16) Number of Significant Parenting Parameters 4 2 3 1 * Parameter estimate at least twice its standard error